Some of us have given a fair amount of thought to this issue and would welcome a chance to participate in a discussion on how meshes might be integrated into SL ... whether or not meshes receive priority over prim modeling tools.

The building community can help in having those with no mesh-tool backgrounds learn how to create them.

Do not lower the bar for building any more. It's wonderful that everyone can build within minutes of entering SL, but they can already do that.

The lower the building bar the lower quality and higher quantity of crap that comes out. If you improve the existing building tools, all you'll do is encourage more prim spam. Allow mesh import and the people who have the ability and have taken the time to hone their skills, will leap at the chance to use actual professional grade tools. Plus, you (LL) seem so keen on SL becoming a viable software/business/commercial/media development platform. This advances all of those and adds design to the list.

Some people have already made SL their lives. For those talented people who haven't, the number one reason is that SL is a waste of time. Even after having created something amazing in SL, they can not turn around and apply the technical skills they've learned to the real world. Using the prim modeling tools is not a marketable skill outside of SL, and only marginally marketable in SL.

From a business point of view it's all about ROI. Improving the prim tools has reached the point of diminishing returns, but the potential of mesh import has only been gently flirted with in the form of sculpties.

My Recomendations:
1) obj file support
2) take the uv map (at least one) from the file
3) Allow a second (simpler) obj file to be used as a collision mesh.
(This can and should be done at upload time, not in world)

One caveat, however - no doubt you'll have to limit polygon/vertex counts, please try to strike a balance between technical considerations and effectiveness. If the mesh import tool limits creators due to excessively restrictive polygon limits, then it will be more of a frustration than an enhancement.

Also, please consider using .OBJ import - including texture coordinates, so we can render out our texture maps and apply them efficiently, instead of breaking them up and applying them singly to groups of objects.

The resulting creativity you'd see from this will be mind-blowing, I can tell you. I've got too many things on my hard drive that I've made that can't be put into Second Life easily because of import/creation constraints.

Please implement mesh import!

On the other hand, I'm leaning towards the opinion that I'd like to see prims continue alongside mesh modeling. New residents should always have an easy way to begin building with no previous experience. It's a large part of what makes SL as approachable as it is.

Of course, if a relatively simple mesh-based in-world building tool could be managed, that would certainly be ideal, even if outside applications remained necessary for more advanced techniques.

We should definitely go towards actual models as the sheer difficulties in expressing detail selectively in a perfectly square mesh just makes it a frustrating and long task of trial and error, hoping that we might actually succeed. I can't stress enough that we should have a variable number of verts instead of the fixed dimension.

Also, please consider allowing custom bump/tangent-space normal mapping so if we can not at least have high detail meshes, we can have high detail bump mapping so as to increase that detail. I'll even make a patch exemplifying that if you want.

Whatever you do, PLEASE do not use COLLADA.... the world does not need to use XML to transmit model data. Try just using OBJ or at least a format that can be directly converted from OBJ with a converter instead of just providing an in-maya tool this time... I know you like community involvement for these tools but a reference implementation would be nice.

+1 to variable mesh sizes, sometimes an object just needs 12 triangles, sometimes it needs many more.

What i mean is that you would be able to upload an obj file up to 512 triangles and have that count as a prim.

If mesh import doesn't happen please take a look at openCroquet for ideas on inworld mesh creation, especially extruding from a curve.

Support for vector graphics would be great also.

To address some of those concerns, I would like to suggest that if you do implement model imports, part of their properties window be a count of triangles. transfer size of mesh could also be useful. Similar displays on groups of primitives would allow for comparison.

You also might consider giving meshes a weighted prim count. For instance. for every X triangles, this mesh counts as another prim. So you might have very very complex meshes that count as up to 255 prims. All the rules for parcel return, linking, and being physical could then apply in proportion.

I believe the argument against weighting before, has been that it might be too hard a concept to follow. However, since it applies to people who have the brains to use a modeling tool to begin with, I don't think it's unreasonable. The math is elementary (school).

However, mesh import will allow me to do FAR more quality content than I can now. As my texturing is tightly integrated with my modelling pipeline, and that SL doesn't allow me to download items I have made in the viewer, I end up taking far too many extra steps just to finally texture something I've made with parametric prims, by literally recreating those prims in the modeller, even though I know I will be making it as a prim in world.

Another thing I want to have looked at is animation and custom uvmaps. I'd really like to create items with distortion free uvs, and currently, I cannot do this easily with sculpties. I'd like to see bones/skinning so we can have items animate about realistically. There is far more room to run I believe if we create the framework in SL to work with external tools.

Voted.

Prims still have their place though...they're there, they're accessable to most. A lot of improvements could come just by a UI overhaul of the build tools, which would fall to a different team and could be worked on without excluding other things. The once-promised "fuse prims together as a mesh" feature could also be more possible as a result of doing mesh support.

Also, official support for prims >10m plz!

Too late, as the world does already. Your avatar shape is transmitted that way. The avatar skeleton is sitting on your drive as an XML document.

COLLADA fits right in, and there is already some support for it in the viewer.

Aaron wrote:
" You also might consider giving meshes a weighted prim count. For instance. for every X triangles, this mesh counts as another prim. So you might have very very complex meshes that count as up to 255 prims. All the rules for parcel return, linking, and being physical could then apply in proportion."

He said it better than I could. This is an excellent idea for how to reconcile .obj meshes with the current prim system.

I can grovel if it will help...obj....pretty pretty pleeeeeease?

There is nothing I would like to see more in SL than this. While I have thrilled at the ability to product organic shapes, I have always felt Sculpties were too limited.

gIVe uS THe MeSH IMPoRT oR ThE BuNNy GETs iT!

One clear advantage for potential content creators is that they can create, even sell meshes completely outside of the SL economy, in the RL economy. This is also really inevitable and, for better or worse, put content ownership and distribution responsibility more on the shoulders of the creators themselves.

I don't exactly know where it would fit, but we should include a discussion about COLLADA. I wonder how many of the attribution and upload issues would relate to COLLADA adoption, but I know little about it specifically.

One of the greatest challenges to bringing meshes into SL will be managing level of detail (LOD). This is already an issue with sculpties and tools have been built (Blender plugins) to help plan for this. The LOD specification would be a key component to making external tools work well with SL.

I would personally also like to see more REST and other interfaces supported by the libsl or other APIs that allow direct upload of assets, with charge or whatever, and the ability to directly cloth them onto one's avatar. I have been hinting about creating Photoshop macros to preview skins and clothing someday with a single click on a plugin or macro rather than going through the several step process today's workflows require.

Thanks Quarl for looking ahead. It keeps us motivated to stay engaged and informed.

in-world tools are fine, and should be improved in the general course of things. Lack of mesh support is a deal-killer on lots of potential use cases. It also kicks sand in the face of a huge community of 3D artists who have honed their skills over the years, and would be an incredible fan base if the situation were to change. Right now they are looking at Qwaq, Forterra, X3D etc., anything they can to get into the game. Linden is a non-starter with inworld-only tools.

http://www.realityprime.com/articles/volumes-of-reading

I gotta think we can do better than what LL does right now, but doing it without raising the bar on the system requirements for the SL viewer is an important consideration also.

How about a compromise? A lot of people want to see avatars improved - why not limit it to uploading a mesh for an avatar? Cap the vertex/triangles limit to that of the current avatar mesh, or maybe a bit higher, and make note that the bounding box for collisions WILL stay the same as current, so that the physics engine doesn't care.

I would assume there's nothing stopping LL from requiring modelers to upload a companion collision detection mesh, or even providing a tool to add a simple caulk immediately after upload. Another possibility: perhaps there's no coldet for uploaded meshes; they're all "phantom".

In the end, it's all about providing choices. Obviously object mesh importation will never be a desirable option for a large portion of the population. It will require experience with modeling and modeling applications (most of which have a staggering learning curve in comparison to Second Life's building tools). On the other hand, there is a portion of the population who already have this experience or who are willing to learn in order to create content in a more exacting way.

Anything that allows for better content without placing excessive limitations on the population is a good idea, in my opinion. I feel that this is one of those ideas. Those who have the skill to import their own meshes will no doubt create exciting, innovative things with it. Those who cannot will continue to use Second Life's current systems for content creation, exactly the way they have been before.

Collaboration, in my eyes, necessitates remixing of content.

Prims are ideal in this respect, because they are easily modified by another party.

Meshes, however, are a little more cumbersome to remix.

My vote is for importing meshes, but would hope that i could also import simple prims from an outside program and they would remain simple prims when i import them into SL. It would be painful to draw a cylinder in Blender or Sketchup up only to have it triangulated or converted to a mesh, once i brought it into SL.

And again, coming from the collaboration perspective, 'import' is only half the story. Being able to export these meshes and prims should also be high propriety. That is, if Second Life fancies itself as a burgeoning collaboration platform and not just an elaborate 'viewer' of content.

I always thought is would be amazing if SL adopted Blender or BRL-CAD as the inworld modeler. It would be a win-win for both SL and these open source communities.

Could someone please give me a quick technical reason why this isn't possible?
Cheers, Theory

************
For import, the desire is to use either existing 3D creaton tools or new tools that might be created. Today, a number of content creators are using Maya or Blender or some home grown scultie tools. This has been very helpful but is still a difficult path.

We need to get to a point where content creators can use there own tools without any hardship in getting their creation into Second Life.
************

So meshes would be a big help for either content creators using professional tools or developers wanting to make thier own creation tools (like what happened with sculpties - see: http://wiki.secondlife.com/wiki/Sculpted_Prims:_Resident-made_Tools )

I realize there must be limits, such as vertex count or face count, and such limits seem reasonable to me. Limiting meshes to a maximum size would be better than limiting meshes to an exact number of vertices or faces.

I do not favor any method that relies on Maya, 3DS Max or Lightwave. I use blender now, because of Domino's amazing scripts, and prefer to use such cheap or free tools.

I do not favor any method that requires NURBS modeling.

I agree that the current sculpty implementation (as of 27 Aug 2008) wastes resources because a sculpty has a fixed number of vertices (or is it faces?). I've made many sculpties that waste vertexes in areas where few were needed.

That's why I'd prefer a "maximum" mesh size (vertex or polygon count, I mean). It'd be great to build less, not more, whenever I can.

Along with mesh import for buildig, we need better avatar meshes, and possibly a way to import other base avatar models with associated morph targets, like loading a new character mesh (David, Victoria, Furette...) into Poser or DAZ|Studio.

A heck of a lot better than sculpties are right now, which don't have accurate collision whatsoever. Collision meshes for simple convex shapes aren't difficult to sort out.

"How about a compromise? A lot of people want to see avatars improved - why not limit it to uploading a mesh for an avatar? Cap the vertex/triangles limit to that of the current avatar mesh, or maybe a bit higher, and make note that the bounding box for collisions WILL stay the same as current, so that the physics engine doesn't care."

That made me laugh out loud. The avatar mesh is on your hard drive, its entirely client side. Uploading avatars will require a complete rethink of the avatar. But in my mind, I would prefer the avatar to remain a client side thing. And collision is even more complicated with avatars... mind.

Why you would think its a compromise to upload a heavy mesh such as an avatar (THOUSANDS of vertices) and the amazing lag that would cause, because not only that, the avatar uses morph animation for the face, which makes for thousands of MORE vertices (every expression - requires another head mesh - to simplify the explanation), as well as being a skinned mesh for animation - how this is better or some sort of compromise over a simple import of small meshes under 1024 vertices (keeping them equivalent to the weight of a prim) only seems to illustrate to me that you have no idea what you're talking about. :/

And I just dont want to hear completely uninformed people screaming they can't upload Victoria 4... I just don't.

Avatars are really unrelated to the discussion... and need to be concentrated on in their own JIRA, as they have unique needs with animation.

Seems to me that this issue is objectionable for some folks from the desire to hold talented people down to protect their inworld business, rather than thinking about striking a balance between what is best for both content creators and users of the service - performance needs to be THE foremost factor, not a fictional equality between mesh modellers and prim modellers. Truth is, prims should not go away and are still going to be a useful tool in any builders repetoire. I believe they need expanding, especially in the areas where they excel.

For the simple fact that they are efficient and small to transmit over a network, and when you get to larger polygon sizes... parametric prims are a better way to go.

Protectionism doesn't work in the real world and it wont work in the virtual one either. Do that and you are kissing SL to a competitor who will take ALL your customers to a new world down the road. That to me is far worse a fear than learning how to make a few simple meshes in a free modeller such as XSI Mod Tool, Wings or Blender - which are perfectly servicable for the job when you need it and used by other game modders for years, not pros.

Anyone who's seen OpenCroquet has seen within is a means to make meshes 'in-world'.

You could be an hero by replicating this functionality in SL (and receive many cuddles from exotic creatures).

In the context of SL, a system which uses prims to create vertexes and allows welding and interactive editing by multiple parties would be cool.

So, while you are working on the prim editing tools, create a mesh editing facility which is integrated, and allows for the creation of meshes which are uploaded directly to the asset server, or downloadable as your choice of ham, bacon, or nougat.

So, add yourself a new 'prim type', which is 'mesh vertex', and allow 'mesh vertexes' to be 'linked' to other 'mesh vertexes', and have the viewer draw the points in between each mesh vertex, etc etc etc.

If you have questions about this evil scheme, talk to V'Ger, or IM me.

Other people type too much. Don't listen to them. Next they will try to tell you that I am just a subplot.

Enabling COLLADA import will allow enabling completely new toolchains of data to come in. Data from professional modeling tools, Data from the CAD space, Data from outside simulations. It will help Second Life be used by more classes of people, and help solidify Second Life as the development platform for a lot more people.

Please contact me if you have any suggestions for changes to the COLLADA format to support Second Life. I'm in the COLLADA standards group, and we are just planning features for version 1.6.

Many enterprise customers need to pull in data from other sources, and COLLADA was intended to be the interchange format to move data between tools. I get beat up constantly by customers that simply can't cope with Sculpties.

Additionally, please allow all clamps like limits on number of allowed vertices to be server side checks, and not limitations imposed by the viewer. I'd like the Second Life viewer to continue to be common between Second Life users and OpenSim users, and some use cases that OpenSim is used for might have different requirements than Second Life's main grid.

Thanks for putting this up for a survey.

My viewpoint on this topic is slightly different:
if meshes are going to suffer from all kinds of problems (which i don't know of, since i don't know, how meshes would be implemented at the end), i'd rather like to see sculpted prims be optimized instead. I'd rather like to see things like this: "make animated sculpties, solve current inconveniences,allow more precise sculptie meshes..."
But this is not asked for in this vote ;-(

If the votes where:

"enhance sculpties" vs. "make in world editors" (i'd vote for enhance sculpties)
"enhance sculpties" vs. "support for meshes" (i'd ask for more details, see below)

So, in order to vote for meshes, it would it be interesting to know, where we can find more detailed
informations about the anticipated results, so we can back our decision with (even weak) facts?

things like:

will mesh-support be more or less resource consuming compared to sculpties?
will mesh-support also handle animated meshes ?
will mesh-support include texturizing support ?
will mesh suport also mean, we can EXPORT meshed from SL to out tools ?
will mesh support make sculpties obsolete ?
How would sculptie tools look like ? any phantasie about how that could be ?

It would be great to see a matrix with pros and cons for sculpties/meshes...
Then sort of a decision could be made, whether it makes sense to vote for meshes.

But since all that probably won't be available (sigh), i rely on my stomach now, and it tells me (without knowing any facts at all): Vote for meshes (and hope for a .OBJ importer)

So please, if you're going to let us import meshes, use one of the hundred existing formats instead of inventing another proprietary one that will require specific modelling software with a mess of custom scripts.

As I said before, the REAL issue for nice appearance in SL, as far as I know, is network bandwidth, not vertex-count. And, Qarl just added the upload/display support for the new x by y sculpty textures. Before I'd support new types of meshes, I'd like 1) to see what can be done with the new sculpties, 2) to see what the original programmer left out of the primitive set and options that SL currently supports, and 3) see what the bandwidth cost of supporting new meshes will be like. Oh, and better caching of textures would be a good thing to implement first, as well. Not to mention better animation support for sculpties, and stitching support for scultpies of various sizes and shapes.

My guess is that for the kind of virtual world Second LIfe is (arbitrary content used by arbitrary individuals in arbitrary simulators), ad hoc geometry like scultpies and primitives will ALWAYS look and behave better until everyone using SL has a direct T1 connection all the way from the sim to their home computer.

And no, in this case, the fact that YOUR mileage may vary doesn't matter one bit: it's the average case that matters, not the extremely fortunately few who currently have such connections.

Vertex count is directly related to mesh size which is directly related to bandwidth. We should not have the ability to import large meshes into SL, end.

Performance in the viewer is a combination of vertices viewable on the screen (directly impacting frame rate) and bandwidth considerations.

Small meshes will be highly useful for NON quad topologies, something that is not very feasable with sculpties and prims. (not impossible, just very prim AND polygon intensive if you use a prim to draw each triangle, or highly annoying and not very high quality to do with sculpts in comparison to using the actual meshes.) I have done them ALL, and so I can compare. When you have done them all and can compare, then you are in a position to talk about it.

I do a lot of this kind of modelling, so I am very MUCH in a position to comment about it from a point of experience. I'd like to be able to use my own gemstone designs in SL. They are VERY small meshes. I can't very easily. This annoys me to no end.

I was the one who posted the edge stitching and sculpty skinning animation JIRAs. I still want these improvements to sculpties to enter SL. Sculpties are highly optimised for NURBS and subdivision surface modelling - they should continue to be improved. And let us use something else for hard edged low poly objects with non-quad topologies. (using triangles instead of quads) Sculpts were NOT designed for this kind of modelling - its trying to fit a square peg into a round hole to keep on bothering.

Use a hammer when you need a hammer, and a drill when you need a drill. Let's stop using a hammer when we need a drill.

VWR-5189 and VWR-5191 are the edge stitching and sculpted prim rigging jiras, respectively.

I read Avi's article, and again, he makes the same point many of us make when supporting low poly mesh. Prims are not automatically better from a performance perspective as they are wasteful with polygons. This problem impacts frame rates in the viewer.

"The downside, of course, is that these pseudo-solid volumes are not as efficient to render as arbitrary hand-crafted polygons, since there are lots of overlapping and hidden surfaces (which take time to render, even if you don't see them) and require lots of tiny state changes, including simple changes in position and orientation, each of which takes a small-but-cumulative amount of time for any 3D hardware system to set up and work through."

http://www.realityprime.com/articles/how-sl-primitives-really-work

This is the program I use to make gemstones - I have used it for years. (way before SL)

http://www.gemcad.com/

Note its not a "3d modeller" per se. It is a lapidary simulator, for cutting gemstones. Most users of this program are not 3d modellers... they are jewelry designers. Most of whom are not very computer savvy. (I've spoken to many when showing my 3d renderings made with this program)

Here's some of my work from way back in 2004, when SL was just a fledgling.

http://www.flickr.com/photos/hypatiacallisto/2812253153/

The center diamond is a replica of the Hope diamond. Want to know how many faces it has? A whopping 96 polygons. Tack on a few more when triangulating the table facets, so if brought to SL, would maybe be... oh, 110 ish or so.

That is ALL.

A sculpty is MUCH larger than this! The OBJ for just my Hope is 25.5 KB. Tiny. A sculpt texture is both much larger in texture and polygon size, as it has to define the placement of 1024 vertices, not the 88 vertices which is all this model comprises! Prims inworld are also much LARGER than this!! It's very hard to get a sculpt to do this kind of nonstandard geometry accurately, and with gemstones, ACCURACY of angles is PARAMOUNT when you are doing replicas of actual stones for a jewelry design.

http://en.wikipedia.org/wiki/Image:Inside_Beijing_National_Grand_Theater.jpg

... or similar buildings. There's only one catch: The client wants you to build it with no more than 30 prims.

Now imagine Second Life with mesh import. The surface of a sculpted prim consists of 1024 quads (2048 triangles, 32x32 vertices). 30 sculpted prims are equivalent to a mesh consisting of more than 30,000 quads.

Just to get an idea of what can be done with 30,000 quads in a custom mesh, check out these pictures:

http://img329.imageshack.us/img329/6485/picture6zd0.png
http://img329.imageshack.us/img329/5486/picture5lv2.png
http://img404.imageshack.us/img404/6523/picture4xd4.png

This structure was modeled in Blender, a free 3D program. It consists of 29,960 quads. Rezzing it in SL should cause no more lag than 30 sculpted prims. Building a similar structure with prims would require 2480 hollow cubes, one per window. It could be done with a script, but the result would have ugly seams all over it due to prim alignment issues. Making this in Blender is a matter of minutes. The result will be seamless and use fewer polygons than the prim-based version.

Mesh import will liberate builders from the restraints imposed by prim modeling. We will be able to achieve more in shorter time frames and with fewer resources. If mesh import can be implemented, there is really no valid excuse for not doing so. It will improve the platform tremendously.

Having said that, as a beginner, development of mesh support would seem to be the preferable route since it would not require learning to handle "sculpts" here and "meshes" in most other worlds, real and virtual, where 3D tools are used, especially since the eventual goal is to be make this grid a more connected part of the larger community of virtual worlds.

"Liberating builders from restraints imposed by prim modeling" would be wonderful . . . and wouldn't this reduce the load on the SL servers?

I have been creating game assets for 17 years. Mesh is easier than sculpts. Mesh can be quicker to produce,. and is an additive process. Creating mesh is open to a lot of cheap and free 3D tools already extant, so rather than limiting content creation to a few "pros", it actually opens up wider avenues of asset creation, with a variety of 3d tutorials alreay available on the web. and finally implementing this will remove a major bottleneck we have in our production at THI, being the creation of acceptable, professional level textures. The current texture tools for Sl are entirely inadequate, resulting in painfully tedious methods (see: http://tumansky.com/blog/2008/08/how-to-texture-in-sl-like-pro-part-1.html ) taking about 200 man hours to produce textures that could be done using standard UV editing and photoshop, about maybe 16 hours or less. Just make sure to have the file type be something extant, and common to as many 3d programs as possible like *.3ds or *.obj, or even *.ASC

The mesh itself can be more efficient that the current prims, and as an airrcraft builder the problem with crossing sim borders is most noticely caused by a high number of attatchment or vehicle prims crossing the border with the associated Avatar. Also attributes like hard or soft edges, and different surface attributes can be applied by the face by material ID. if the Polygon count is low enough it could have true collision like the prims rather than the approximated collision of sculpties

Limitations to preserve bandwidth, or keep the assets small for storage can be implmented in ways similar to how SL handles files outside acceptable parameters in other file types such as sounds and animation. You can set a hard limit of how many polygons or vertices that can be uploaded. You may want to go with verts, because each attribute, such as the UV assignments and material IDs will add another layer of verts to the existing ones. The average Game character is between 4000 -7000 polygons. If LL were to limit the number of polygons to 4096 per object, for example, and fail anything over that amount. Someone above posted the number of polygons in a sculptie. Keep that as the bench mark. You could further limit the mesh to only allowing mesh that passes a Stereolithography Test (STL Test) so make sure the mesh is properly closed , welded and legal, and fail to upload it if there are mesh errors. A high upload cost might or might not be a good solution, as I see it causing a bit of inflation in world as yhe greter cost to upload is passed on to consumers, but 10 or 20L would work. But a further limitation may be to have those loading Mesh also have to hand optimize two more Level of Detail (LOD) meshes of lower polygon counts, for distance or performance limited clients. Also the upload should fail of the object loaded has a bounding box larger than 10m on a side (conversely all objects rez in world in proportion but at .5m on the longest dimension of it's bounding box). I would suggest against "fixed" point compression of mesh as that brings in imprecision that causes people to tear their hair out. (pre-SL Uploader Sculpties looking melted. or abused)

A blue sky would be able to bind a user created mesh to the SL animations and skeleton (So one could fix the UVW errors in the default male and female, or create a new AV type). Less blue sky would be to allow for some mesh morphing or bound motion similar to the Mesh Linden Trees. or some export function to allowfor the addition of Flexi attributes. But I would be happy to get mesh into SL proportionately with good UV maps .

To Conclude, yes I'll vote for this, as I think it's an excellent idea. We (THI) really dont' use standard prims any more, using almost exclusibvely sculpts these days, so those tools would generate limted enthusiasm. Comments?

--Karl

P.S. making them unable to be duplicated by copybot2.a would be an added plus.

I have this feeling that allowing meshes, in a common format such as OBJ, can only improve both the quality and variety of content. Like it has been said before, there is a variety of tools available , ranging from simplistic free 3d tools, to industrial monsters, including real good opensource ones such as blender. Most of these allow anyone to create OBJ files. Everyone can quickly start creating meshes even on little systems, and skilled people can achieve wonderful things. OBJ files are to 3d what tga files are to 2d, I think. I don't see a future in 3d virtual worlds without including these. Even more if I think about interconnected worlds. ( eyes suddently shining )

I think prims have some advantages : you can create them inworld, starting with zero 3d knowledge, by using inworld tools. This has allowed thousand people to start creating 3d content, and this is a good thing. Prims also natively carry performance information, such as LOD. They can be really lightweight on the bandwidth. They certainly contributed to the crazy development of SL too.

They have a big drawback. You have to develop insane amounts of tricks and cleverness to achieve certain results, often ending in huge linked sets, that completely kill the performance aspect I mentionned before. Quality content creation is made quite difficult in my opinion. You spend hours to go around limitations, your brain is focused on technical questions, when you would need to be able to let your creativity go fluently through the interface. I think sculpted prims allowed to approach this a little, but still you have to figure out quite complex things if you want to make really good, clean ones. And then, some shapes are just impossible to create, due to the topology limits. New sculpted sizes will expend the range of possibility, but they also add a new level of complexity in the process of creating. In most cases I end far from what my idea was in the start. The tool is acting against the artist creativity, I think.

Creating meshes is straightforward. With most mesh manipulation tools, when you create a mesh, your spatial sense is just free to express, you don't need to split your head between your art and some obscure mathematical alchemy.

Then, we need efficient meshes, so that performance is not killed. Begginers will tend to produce unefficient and heavy meshes. More experimented people will know how to reduce the load. Its hard to figure out how this can balance in the end. A 6 vertices box is certainly lighter than a prim box, I think.

OBJ files can carry UV maps. To me this is a big lack in the current SL implementation. This would allow some really efficient quality texturing, when prim texturing is a sliders and numbers nightmare. I have this sweet dream of a SL with bump, normal, specularity, alpha, mirror...etc...maps.

I have been throwing ideas as they come, but I think in the end I would say :

*keep prims tools, eventually improve. They allow people with zero knowledge to create. They are lightweight.

*allow OBJ import, restricted to a max vert count. Maybe create something to encapsulate LOD meshes in an object. allow UV map import. Go interoperable and standard.

voted

In actual practice, that's precisely what sculpted prims are often being used for - with widely varying degrees of success. While nicely rounded organic shapes are nice, people really really want a way to create accurate arbitrary shapes.

As much as sculpties caused a tremendous buzz and comments such as "this will revolutionize building", they are a mixed bag : Very disappointing for many of us because of the dramatic shortcomings for making non-organic shapes, but still worlds above regular prims for many uses.

Part of its beauty is its relative simplicity, and trying to satisfy 'standard' requirements for a 3D application is within the scope of SL, but I think not in the immediate future.
In the wider world of serious 3D apps we are now approaching a relatively standardized interface/gizmo/terminology set for 3D apps in general. So maybe the day isnt far off where implementing a comprehensive 3D tool interface into a virtual world viewer is relatively trivial... but the general user has to intuitively understand how to use or ignore it , and that may not be quite so close.

Forcing the user to convert to the current simple RGBA image format solves a multitude of problems.
Primarily it forces the user to understand the limitations of the medium, and adapt their model to fit as perfectly as possible within that necessarily finite system. (data packet sizes, rendering power, etc)

I vote first for "mesh import" (MISC-1494), because its what takes us immediately forward, and allows us to use the tools of our choice. Which will be significantly more powerful and refined than any inworld tools for some time to come.
I also vote for "inworld tools" (MISC-1495), because as a natural course of events this will naturally evolve out of the previous issue in a more timely and on-demand manner.

MISC-1494 = important now
MISC-1495 = will come from MISC-1494 anyway.

on a second note for those who don't know how to model that well. I do also agree on having better tools for prim shaping.

• Improved performance because of more efficient builds
• Potentially lower texture memory consumption, because custom UV maps allow for more efficient texture usage
• Rezing of meshes can be controlled better than the somewhat faulty sculpt mechanism (it is odd to see lots of spheres with fully rezed 1024x1024 textures. something is awry there..)
• greater freedom of creativity
• an even higher potential for outstanding builds

I have lots of ideas that would be easily realized by mesh importing, and it would benefit SL on so many fronts. Already it is impressive what sculpts did for SL, but I believe the system is too ineffective in its nature, by constraining to 1024 vertices per prim. Im not necessarily voting for more vertices, but for the option to use LESS actually. sometimes you dont need 1024, but you cant save on performance at all there. Also, UV maps for sculpts are cumbersome to handle, custom UVs would be easier to manage and can be more effective, by enabling things such as texture sheets, as in, condensing all needed texture data for a build into 2-3 large sheets, instead of dozens of small textures

I also have a couple of concerns:
1 - Will there be limits to how detailed the meshes are?
I really do NOT want to see avatars wearing 10 meshes each with 50,000 polygons.
I DO want to be able to make a relatively simple mesh with perhaps only a couple or several polygons (think two sided meshs, it would cut down on triangle usage of prim plants even more).

2 - Will we be allowed to use .obj meshes?
I, like many others can't afford to spend hundreds, on specialised software just so I can use a specific format. I use free software mostly (Wings3D for sculpties). We need something that 99.9% of SL users can use, even if that means LL having to support several different types.

If i understand you correctly...

I also have a couple of concerns:
1 - Will there be limits to how detailed the meshes are?
I really do NOT want to see avatars wearing 10 meshes each with 50,000 polygons.

Well, somebody else made the point that even if you do limit mesh density, there is an easy way around that (linking multiple low poly meshes together to make one high poly mesh). So, while I do think that is a moot point, what you should consider is that if someone does happen to make a bajillion-poly neckless, nobody is going to buy it, or even want to wear it for that matter. Perhaps a limit isn't out of the question, but something people need to realize is that it wouldn't enforce the level of control that you would, at first, think. I do also realize that this could open the door to some new griefing opportunities... but maybe LL could give residents the ability to limit the amount of these objects (or even just the poly density) being displayed on their own screen? You could, maybe, set the limit of what's being displayed yourself, in accord with what you're particular machine can handle. There is already a "mesh detail" option slider in the graphics tab.

2 - Will we be allowed to use .obj meshes?
I, like many others can't afford to spend hundreds, on specialised software just so I can use a specific format. I use free software mostly (Wings3D for sculpties). We need something that 99.9% of SL users can use, even if that means LL having to support several different types.

Most times, it is up to the developers of a software suite to enable support for a particular format. Supporting several formats is backwards. For example, Blender is free, and whether the residents want to take the time to learn how to use it is not indicative of whether or not they CAN use it. Of course they can, everyone can. Blender does not discriminate. Limiting all residents to weak software is also backwards.

Also, whichever format LL decides to incorporate (as I doubt they would spend the resources to develop their own), I'm sure that there will be a plugin released, for whichever software suite you happen to use, supporting this format, soon after any anouncements... if there isn't one already. That is what is so great about the community

Why would you want to edit the mesh after it's already in world? Why not get it right before you upload to SL... that sounds like the sane alternative, compared to writing a bunch of functionality that isn't really a necessity.

In the words of Quentin Tarantino on "Pulp Fiction":

"I didn't need to make a director's cut, I made the movie I wanted to make the first time around.

Also, a common misconception I'm seeing here is that people think .obj is the answer to everything...
Well, lets take a trip down memory lane. Obj was developed by Wavefront way back in the early '90's. Even before the SGI merged them with Alias, and way before Autodesk purchased the offspring. This predates most of the software utilities out there, and it predates SL by like 10 years.

My point is that .obj is old, and all it accounts for is geometry and UV texture space.

If I could get everything I wanted, I would have LL incorporate support for a different standard.. COLLADA.
My arguement being that COLLADA encompasses animation, geometry morphing, skinning (etc, etc...), not just static geometry. It is also supported by like.. everybody, for like.. everything. COLLADA's entire purpose is to be an open 3d standard for everyone to use.

Check it out:
http://en.wikipedia.org/wiki/Collada

.obj support will just allow the creation of a bunch of new, static stuff, thats it. Support for a standard like COLLADA, on the other end, would eventually lead to something like, maybe being able to create your own creatures in a real, not-hacked, non-work-around sort of way. Not to mention the ability to create your stuff out of something other than funny-shaped, virtual legos. And I'm not saying they just jump in and throw all of these features into SL all at once. I mean, lets just start at geometry, but at least using this sort of standard will leave a door to all of these other features, to incorporate down the road some time (instead of having to take a sledge hammer to the wall when the time comes).

SL is constantly growing and expanding, and I'm sure LL wants to move the process on as reliably and quickly as possible. I know a lot of people are arguing against mesh support because it isn't conducive a streaming, sprawling virtual world. But wake up! Maybe something like this will cause lag today.. But what about two years from now? I guarantee that if the wheels don't start rolling on something like this in the present, when it gets to be 2010 or so, everyone will be looking back and kicking themselves in the ass because these features weren't implemented now. Internet speeds will be faster. Computers, on average, will be faster. The state of the art will be able to smoothly handle this sort of thing, and, without it, all you'll notice is the vacuum left by its absence. Not to mention the vacuum created by the absence of all the users that abandoned SL, and went other platforms that will/do support some sort of arbitrary mesh support. Technology advances at an exponential rate!

For example, the second Google allows user generated content on Lively, my staff and I will be there.

I know I know, people won't automatically be able to create stuff with all of these tools using only their intuition. But it isn't exactly like anything is being taken away. Nobody is talking about deactivating any of the standard building tools in SL. All we're talking about here is opening the door to much more powerful tools. And you know what? It isn't impossible. If you're able to build in SL, chances are that if you can get on google and read a few tutorials, you'll be on the road to figuring out even the most advanced 3d suites out there. I guarantee that for every one tutorial out there on sculpted prims, there are 10 tutorials on poly modeling. I cannot, for the life of me, fathom why you would ever want to limit the abilities of everyone as a whole, because the general community doesn't know exactly how to take advantage of them. There is absolutely no reason to be afraid of mesh support.

As far as I can remember, nobody exactly knew how to do sculpted prims when they came out either. But when people saw what could be done with sculpted prims, they sure as hell did everything they could to figure out how.

/end rant

>> The only thing I hope for, if meshes DO get allowed - there SERIOUSLY needs to be some
>> way to edit them in-world. I'd like to see some way to handle sculpt maps too, but ...
>> the point of the sculpt maps was that they were a quick and dirty hack, that required
>> little heavy lifting to get into the system - I get them impression meshes would take a
>> bit more.

> Why would you want to edit the mesh after it's already in world? Why not get it right
> before you upload to SL... that sounds like the sane alternative, compared to writing a
> bunch of functionality that isn't really a necessity.

I think you're missing at least some of the points I and others are trying to make on this.

> "Why not get it right before you upload to SL..."

Some of us aren't looking to upload in the first place. I'd like to click somewhere in world, "Create New Prim", and select prim type Mesh - and from there, create something. I already know the interface to SecondLife - I don't want to learn a whole new interface - one that's usually quite complex.

Which leads to the second problem with that statement:

While I'm editting it in world, my friends are standing around kibitzing, making suggestions that they see in real time. SecondLife is a collaborative environment, not a setting for set pieces of art. People get freebies, and modify them into something new, and while yes, you could download free models to load into whatever program you want to use to edit them (yeah, I've used Blender - and seen a few of the other programs in screenshots - definitely not what I'd like to play with) - but that's still taking you OUT of SecondLife into some other program - back into RealLife (RealLife is Trademarked and Copyrighted by the Diety of your choice, 13 Billion BC)

People want to make SecondLife more realistic - well, then we need to quit dragging ourselves out of it to do something that is meant for in it. I'd like to see basic texture editing tools in-world as well - something to change color balance, gamma, and edit pixels at a basic level. I don't really want PhotoShop-level features - but it'd be nice to be able to rotate the color of a blue-tinged texture to green, or red, or to fix that oopsie where the scanner left a couple of spots of white from dust on the platten.

> In the words of Quentin Tarantino on "Pulp Fiction":

> "I didn't need to make a director's cut, I made the movie I wanted to make the first
> time around.

But SecondLife is much more collaborative than movie-making. There is no one "Director" for SecondLife - there's people working together, which is what the original vision of what SL could be, was.

> Also, a common misconception I'm seeing here is that people think .obj is the answer to
> everything...
[...]
> My point is that .obj is old, and all it accounts for is geometry and UV texture space.
> If I could get everything I wanted, I would have LL incorporate support for a different > standard.. COLLADA.

> My arguement being that COLLADA encompasses animation, geometry morphing, skinning (etc,
> etc...), not just static geometry. It is also supported by like.. everybody, for like..
> everything. COLLADA's entire purpose is to be an open 3d standard for everyone to use.

And that may be exactly the reason everyone is pushing for .OBJ and not for COLLADA - they want something that can be used now, and not something that will end up taking forever to fully implement. I haven't read the WikiPedia article - I suspect COLLADA is built in an extensible fashion with subsets of limited functionality that can be implemented in blocks - but then SL has to handle the huge list of calls from people trying to import some model they found for free on the Internet that uses some part of COLLADA that they haven't implemented yet.

> .obj support will just allow the creation of a bunch of new, static stuff, thats it.
> Support for a standard like COLLADA, on the other end, would eventually lead to something
> like, maybe being able to create your own creatures in a real, not-hacked,
> non-work-around sort of way. Not to mention the ability to create your stuff out of
> something other than funny-shaped, virtual legos.

But that leads to still MORE complexity in creating the content. And that complexity means more trouble for the client end.

> SL is constantly growing and expanding, and I'm sure LL wants to move the process on as
> reliably and quickly as possible. I know a lot of people are arguing against mesh support
> because it isn't conducive a streaming, sprawling virtual world. But wake up! Maybe
> something like this will cause lag today.. But what about two years from now? I guarantee
> that if the wheels don't start rolling on something like this in the present, when it gets
> to be 2010 or so, everyone will be looking back and kicking themselves in the ass because
> these features weren't implemented now. Internet speeds will be faster. Computers, on
> average, will be faster. The state of the art will be able to smoothly handle this sort
> of thing, and, without it, all you'll notice is the vacuum left by its absence. Not to
> mention the vacuum created by the absence of all the users that abandoned SL, and went
> other platforms that will/do support some sort of arbitrary mesh support. Technology
> advances at an exponential rate!

You put a lot of faith in the "technology advances at an exponential rate" argument - and I hold that it is flawed - in that it is not a smooth curve. We've hit some road blocks in the computing field now - hard drives keep getting bigger, certainly - but they haven't been getting faster. The hard drives (MFM and RLL technology hard drives) I had on my old AT&T 6300 (8Mhz processor) had a rotational speed of 3000RPM, and a track-to-track seek time of perhaps 20 ms. Now, the best hard drives rotate at 15,000RPM, and have track-to-track seek times of 2 ms. That's nowhere near the factor of 500 that the processor has sped up. Memory is similarly constrained - the main DRAM of the computer is far far slower than the CPU, and so we end up with 2 and 3 levels of cache between, each causing their own hiccups when those caches miss and have to hit the next level up.

Video cards are likewise hitting a wall. The GPUs in the newer video cards aren't significantly faster than they were - they're simply trying to parallelize more. And not everything can be parallelized. We're going to need a jump to a new paradygm - ray tracing may create one for video cards, but then we're going to have to go through a whole new process of optimization, both on the hardware and software sides.

Mesh support is all well and good - but take a look at the fact that this last release of SL caused a lot of computers previously running it to display a message for the first time: "Your computer does not meet the minimum suggested requirements for SecondLife" - I know, as I'm using two of them. And the code to add shadows is moving through QA right now - and requires (last I saw) an NVidia 8xxx GPU, or the ATI equivalent. That's going to knock a whole new group out of the running. Your "vacuum" of people that abandoned SL might not be the ones who left for this hypothetical new world that supports arbitrary meshes - it might be the ones who got left behind and decided that the siliconerati just don't want them. LL is riding a fine line between adding new features - and leaving behind the bulk of people who don't upgrade their computers every time they turn around - because they're the ones who just want to USE them, and not to play with them all the time.

Likewise, network speeds - modems capped out at 33.6K (and then someone figured out a cheat to get 56K asymetric) - and then there was the jump to DSL - and what fraction of the DSL-using population still has the basic 1.5M/256K DSL that was originally introduced 15 years ago? How much of the US is wired to allow DSL2?

Unlike you, I don't have faith that the Internet will just keep getting faster - in part because for most people, and most purposes, it's already fast enough. Hell, how much boost does turning on multi-threading actually gain in SecondLife - and multi-core processors are fairly common now. (In part because really making use of multi-core will require serious refactoring of the code base, and in many parts, completely different algorithms.)

> For example, the second Google allows user generated content on Lively, my staff and I
> will be there.

This speaks of something interesting - you mention your "staff". That implies a business perspective of SecondLife. And while there is a business case to be made, there's also a case to be made for the millions of people who just want a place to visit and build in, who find WoW and EQ and other MMORPG-of-the-hour too violent, too frenetic, or otherwise too limiting for their imagination.

> I know I know, people won't automatically be able to create stuff with all of these tools
> using only their intuition. But it isn't exactly like anything is being taken away.
> Nobody is talking about deactivating any of the standard building tools in SL. All we're
> talking about here is opening the door to much more powerful tools. And you know what? It
> isn't impossible. If you're able to build in SL, chances are that if you can get on
> google and read a few tutorials, you'll be on the road to figuring out even the most > advanced 3d suites out there. I guarantee that for every one tutorial out there on
> sculpted prims, there are 10 tutorials on poly modeling. I cannot, for the life of me,
> fathom why you would ever want to limit the abilities of everyone as a whole, because
> the general community doesn't know exactly how to take advantage of them. There is
> absolutely no reason to be afraid of mesh support.

> As far as I can remember, nobody exactly knew how to do sculpted prims when they came
> out either. But when people saw what could be done with sculpted prims, they sure as
> hell did everything they could to figure out how.

What will be taken away is the ability to point to something and say "Can you show me how to make that?" I remember, about 3 days after I started in SL, I stumbled across someone much older (SL-wise) than I. We got to talking about building prims, and she had a display showing the RSS feed from Slashdot, which was something I hadn't seen or heard of. We talked about building a bit, and she mentioned making micro-prims, and showed me - right on the spot - how to do it.

As for sculpted prims - actually, when they came out, EVERYONE discussing them knew what they were meant for - organic shapes. It was known that they weren't going to be the right answer for making sharp edged objects, but for making apples and bananas, they were excellent. And adding them was a relatively easy project (I think Qarl was one of the leads in doing it, in fact), not meant to be a be-all solution. It's only later than sculpties did come to be seen as a be-all solution to things, and then people started complaining about not being able to make a sharp edge, and all sorts of other complaints.

As it happens, I've actually seen an in-world sculptie maker. It rezzes a bunch of balls and lines you move in the editor, then sends the location data to a web server which e-mails you a texture to upload. Why not, when the prim type is sculptie, just let the in-world editor move the nodes of the sculptie? Likewise for meshes, if/when they come? I'm not saying let's not look at adding meshes to the mix - I just think we need to consider making them fully a part of SL, and not a second-class prim type like sculpties are now.

/No rant, just some points I'd hope will be considered.

Ignoring that much of your argument rests upon some fear you have of learning a new interface, even though said interface would probably provide power and control over your creation well worth the time taken to learn it...

Let's examine what you are saying. You would like for LL to develop their own poly modeling system/software, and surgically graft it to the SL viewer, which, by the way, is already bursting at the seams and quite complex in its own right. So, let's just say that LL won't have any problem developing this monster, and that implementing it within the viewer will be just a breeze (both of these are very naive assumptions, but whatever). You will still run into a wall when it comes to usability/expandability. What I mean is that whatever LL comes up with for these tools, they will never be good enough for everyone (if anyone).

Each different 3d software suite out there has its own arsenal, so to speak, and most of these suites give their users the ability to add to this arsenal whenever they want, in effect, by adding to that suite's core functionality. Not to mention that there are large communities already out there for whatever software you happen to choose, filled to the brim with all these preexisting tools (as in, we won't have to wait around for a large community to develop for SL's new in-world tools) And, I'm talking about power far beyond that of SL's scripting system. Maya, for instance, allows its users the ability to do whatever they could possibly come up with, all within the comfort of two separate scripting languages, not to mention the C++ API. This sort of expandability is not uncommon in the 3d software world.

My point is that not only will simply adding support for 3rd party software be easier to accomplish than adding in world tools (assuming these in world tools are anywhere near equal in strength, again, a very naive assumption), but you will satisfy a larger slice of the community in one move.

"While I'm editting it in world, my friends are standing around kibitzing, making suggestions that they see in real time. SecondLife is a collaborative environment, not a setting for set pieces of art. People get freebies, and modify them into something new, and while yes, you could download free models to load into whatever program you want to use to edit them (yeah, I've used Blender - and seen a few of the other programs in screenshots - definitely not what I'd like to play with) - but that's still taking you OUT of SecondLife into some other program - back into RealLife (RealLife is Trademarked and Copyrighted by the Diety of your choice, 13 >Billion BC)"

Support for 3rd party content creation won't keep you from collaborating with friends. It won't even keep you from collaborating with friends in SL. I keep up conversations in SL and work outside of it in other software all the time.

Also, if you enjoy the collaborative work environment that SL provides, you should look into Caligari trueSpace.

From their website: Caligari has enhanced the modeling, surfacing and rendering capabilities of trueSpace, and the latest version trueSpace7 allows all aspects of real-time design, modeling and animation within a virtual 3D space shared by remote participants over the broadband internet. The trueSpace7 collaboration server enables multiple participants to connect to a shared 3D space to create and manipulate shared content in real-time.

Oh yeah, it does all of that, and it is FREE FREE FREE!(&$(&

http://www.caligari.com/

"People want to make Second Life more realistic - well, then we need to quit dragging ourselves out of it to do something that is meant for in it. I'd like to see basic texture editing tools in-world as well - something to change color balance, gamma, and edit pixels at a basic level. I don't really want PhotoShop-level features - but it'd be nice to be able to rotate the color of a blue-tinged texture to green, or red, or to fix that oopsie where the scanner left a couple of spots of white from dust on the platten."

Wishful thinking.

"But SecondLife is much more collaborative than movie-making. There is no one "Director" for SecondLife - there's people working together, which is what the original vision of what SL could be, was."

I was referring to the fact that you are the director of your own project. Really, I'm surprised that I am responding to this, but whatever.

"And that may be exactly the reason everyone is pushing for .OBJ and not for COLLADA - they want something that can be used now, and not something that will end up taking forever to fully implement. I haven't read the WikiPedia article - I suspect COLLADA is built in an extensible fashion with subsets of limited functionality that can be implemented in blocks - but then SL has to handle the huge list of calls from people trying to import some model they found for free on the Internet that uses some part of COLLADA that they haven't implemented yet."

Collada can be used now, just as readily as some sort of direct .obj support. I don't understand why you think that it can't be. What I meant was that you could implement support for collada geometry immediately, and then, down the road, when the time came, you could implement support for animation, etc. I think you misunderstand me, in that collada is a means of conveyance from one platform to another, not a manner in which you have to create your content, or even the tool with which you create said content. Also, if you're trying to upload an animated character when there is only support for geometry, well, shame on you.

"But that leads to still MORE complexity in creating the content. And that complexity means more >trouble for the client end."

No, it doesn't. Functionality, sheesh.

"You put a lot of faith in the "technology advances at an exponential rate" argument - and I hold that it is flawed - in that it is not a smooth curve. We've hit some road blocks in the computing field now - hard drives keep getting bigger, certainly - but they haven't been getting faster. The hard drives (MFM and RLL technology hard drives) I had on my old AT&T 6300 (8Mhz processor) had a rotational speed of 3000RPM, and a track-to-track seek time of perhaps 20 ms. Now, the best hard drives rotate at 15,000RPM, and have track-to-track seek times of 2 ms. That's nowhere near the factor of 500 that the processor has sped up. Memory is similarly constrained - the main DRAM of the computer is far far slower than the CPU, and so we end up with 2 and 3 levels of cache between, each causing their own hiccups when those caches miss and have to hit the next level up.
Video cards are likewise hitting a wall. The GPUs in the newer video cards aren't significantly faster than they were - they're simply trying to parallelize more. And not everything can be parallelized. We're going to need a jump to a new paradygm - ray tracing may create one for video cards, but then >we're going to have to go through a whole new process of optimization, both on the hardware and software sides."

Technology advances at an exponential rate. Have you seen response times for solid state drives? They won't be prohibitively expensive forever. Memory costs are constantly dropping, and their speeds are constantly increasing. I'm really surprised that I am also arguing these points as well.

"Mesh support is all well and good - but take a look at the fact that this last release of SL caused a lot of computers previously running it to display a message for the first time: "Your computer does not meet the minimum suggested requirements for SecondLife" - I know, as I'm using two of them. And the code to add shadows is moving through QA right now - and requires (last I saw) an NVidia 8xxx GPU, or the ATI equivalent. That's going to knock a whole new group out of the running. Your "vacuum" of people that abandoned SL might not be the ones who left for this hypothetical new world that supports arbitrary meshes - it might be the ones who got left behind and decided that the siliconerati just don't want them. LL is riding a fine line between adding new features - and leaving behind the bulk of people who don't upgrade their computers every time they turn around - because they're the ones who just want to USE them, and not to play with them all the time."

SL viewer already displays deformable, animated meshes. It's called the SL avatar. Windlight was a drastic improvement to the rendering engine. Shadows are not (and probably won't be) mandatory in Shadow Draft, it is an option to use if your system can handle it. I actually covered this in the October issue of Castle Keep Magazine (www.castlekeepmagazine.com) I'm sorry that you don't like your computer becoming obsolete over time, but neither do the rest of us.

"Likewise, network speeds - modems capped out at 33.6K (and then someone figured out a cheat to get 56K asymetric) - and then there was the jump to DSL - and what fraction of the DSL-using population >still has the basic 1.5M/256K DSL that was originally introduced 15 years ago? How much of the US is wired to allow DSL2?
Unlike you, I don't have faith that the Internet will just keep getting faster - in part because for most people, and most purposes, it's already fast enough. Hell, how much boost does turning on multi-threading actually gain in SecondLife - and multi-core processors are fairly common now. (In part >because really making use of multi-core will require serious refactoring of the code base, and in many parts, completely different algorithms.)"

Yes it will, it already is. I don't know why you don't have faith in a faster Internet, you already agree that it has been getting faster and faster over time. And, if you're an American, and you think Internet speeds are fast already, know that we are ranked like 9th in the world as far as broadband speeds go. The average speed for broadband Internet access in Japan is 61 Megabits per second. The average broadband speed in the US is 4.8 Megabits per second. Broadband Internet in Japan, on average, is nearly 13 times faster than broadband Internet access in America. Still having faith issues?

You don't see huge performance increases with SL on a multi core system because the software wasn't written to take advantage of multiple cores. Maya/Mental ray were written with multiple threads in mind, and I have seen massive performance increases with the 4 cores of my core 2 extreme cpu.

"This speaks of something interesting - you mention your "staff". That implies a business perspective of SecondLife. And while there is a business case to be made, there's also a case to be made for the millions of people who just want a place to visit and build in, who find WoW and EQ and other >MMORPG-of-the-hour too violent, too frenetic, or otherwise too limiting for their imagination."

Commerce makes the world go 'round. Lively isn't violent.

"What will be taken away is the ability to point to something and say "Can you show me how to make that?" I remember, about 3 days after I started in SL, I stumbled across someone much older (SL-wise) than I. We got to talking about building prims, and she had a display showing the RSS feed from Slashdot, which was something I hadn't seen or heard of. We talked about building a bit, and she mentioned making micro-prims, and showed me - right on the spot - how to do it."

Like I said, in world collaboration won't end with mesh support. I answer questions in Sculpties group chat all the time (even though my business partners tell me I shouldn't help out compitetion ). This won't limit what you can talk about in these instances, it will actually increase the amount of topics you can go over.

"As it happens, I've actually seen an in-world sculptie maker. It rezzes a bunch of balls and lines you move in the editor, then sends the location data to a web server which e-mails you a texture to upload. Why not, when the prim type is sculptie, just let the in-world editor move the nodes of the sculptie? Likewise for meshes, if/when they come? I'm not saying let's not look at adding meshes to the mix - I just think we need to consider making them fully a part of SL, and not a second-class prim type like >sculpties are now."

While that is cool... does the in world sculptie maker allow you to extrude surfaces? How about reverse geometry, paint along the shape's surface or bake textures? Probably not. 3rd party software would allow these and like 9237498273472398575735 other tools.

/end response

I have changed my vote for better in-world tools for the time being, until the Lindens can provide more information on the importing of meshes. So far it appears that most people want this to be aimed directly at professional builders, leaving out people that just want to build random things for fun.

And on the note of broadband speeds, my average throughput is 1.7Mb. It has been roughly that speed for the past 4 years. I have no faith in seeing even 4Mb within another 4 years .

When you're using Wings 3d, you're poly modeling! #*()$&)$

Mesh support would allow you to use tools, like wings 3d, both more freely and more accurately.

This isn't aimed at professionals, it just happens to include them. Most of the tools you're already accustomed to using for/with sculpted prims already do polygons.

The reason broadband is slow in the US is because of the 4th Reich (Comcast, and companies like it) is keeping all of us here in the Internet dark ages.

If somebody wants to model in-world, then fine for him/her, but let the people who don't need to show real-time modelling work how they feel comfortable with, like actual 3D modellers!
I'd just like to be able to use a 3D mesh I have done before, both in SL and out of just SL, so I don't have to rework on everything. I'm sure somebody will be able to make an in-world 3D mesh generator, but PLEASE, let us upload 3D geometry!

I have been looking for this since I started building in SL, because, I hate how a 23000 polygon object, made with loads of unused primitives, that has seams everywhere, can be simplified in a single-textured object of 1200 polygons that looks completely solid and how I wanted... and not letting us, would be like somebody said before: We would be limited!

About the .OBJ format, I agree with it because, it is an easy-to-export format, you just make your mesh, assign UV coordinates to it and there you go... although, Linden Labs could do like many game engines have: their own model format, so it would be like... Half-Life (SMD), Unreal (PSK/PSA) and maybe Linden Labs (SLM/SLA)? That has the option for (optionally smooth) bone weighting if you want to animate it? I think it wouldn't be a bad idea after all...

What?

>>When you're using Wings 3d, you're poly modeling! #*()$&)$
>>Mesh support would allow you to use tools, like wings 3d, both more freely and more accurately.
>>This isn't aimed at professionals, it just happens to include them. Most of the tools you're already accustomed to using for/with sculpted prims already do polygons.

At no point did I say I was not poly modelling. Perhaps you could enlighten me as to what made you think that? I was merely assuming that SL would use the Collada DOM and was hoping that I could continue to use Wings3D, because at the moment (afaik), wings3D does not support, nor has plugins for Collada. I then gave my personal perspective on the ease of use of Blender compared to Wings3D. I'm sure many others won't have the time, abilty, or even want to learn Blender. I apologize for having a personal opinion. It won't happen again (at least not here).
Also, try not to swear next time.

>>The reason broadband is slow in the US is because of the 4th Reich (Comcast, and companies like it) is keeping all of us here in the Internet dark ages.

Good for you. I live in England, and I can assure you BT is just as bad, if not worse.

Just to clarify, I wasn't cursing.

A sculpty-creator/editor in the viewer should eliminate all of that nonsense, and make sculpted prims practical, and accessable to all of us, not just those who can afford the expensive 3D editor programs capable of making decent shapes, and all the uploads to try and get a sculpt map that works.

The alternative, of course, would be as people have stated here, replacing sculpties with actual 3D shapes, in which case it'd be necessary to allow uploads, though I'd still like to see a decent editing/creation tool in the viewer for those of us who can't afford good 3D programs.

I'm offering my vote on the condition that the promised sculpty creator/editor, or a 3D shape creator/editor in the viewer (possibly as a feature in the editor) is the top priority.

P.S. Until all this can be implimented, bumping up sculpt maps to a higher priority than regular textures would be nice. I'd rather have grey shapes than oddly textured spheres.

Seriously, the current creation tools are pretty groovy and I'd say that most serious content creators I know have already come to terms with the idea that they need to learn a 3rd party program in order to make sculpties anyway. There is a whole generation of SLers out there who has already picked up a copy of Maya or whatever and this would enable them to grow those skills beyond the current limitations. I think ti'd be awesome for everyone involved, not just for LL.

Plus, being able to import obj and such would help content creators from SL port their content more easily to other places, making content creation a much more viable career for most people. I think for many of them, they need this kind of ability to feel good about taking the leap into full time content creation.

It would still be annoying that you make something very accurate and people think you ripped off the model from its original source too...

Also, if it was stolen from a company that has no SL presence whatsoever, I don't think anything would happen honestly. For the past 3 years SL has probably been about 10% Half-Life 2 textures, and nobody has done anything. I cant even count the amount of times I've seen prefabs and items sold with textures from that game, and they give no credit to the original artist (Not that giving them credit would make a difference, since they are just stolen in the first place).

Sculpts however are not particularly easy to work with. There are a number of shortcomings:

• For shapes that aren't round you have to deal with two ends of the object being fairly distorted.
• Texture maps get squashed or stretched depending on the shape, unlike meshes you have no control over this, it's dictated by the shape.
• Sculpts are only an approximation of the original mesh. Where a modeller would have total control over a curve or edge instead they get whatever the sculpt gives.
• Some shapes are flat out impossible to make in a sculpt and have to be simplified or flat out avoided but could trivially be made in a mesh.

Some big advantages of meshes:

• Workflow benefits

• Generally much quicker to work with to accurate reproduction of shapes.
• Much faster to texture. Prims have to have the texture sized and fitted to every face which is extremely slow (There is planar but this doesn't remove the issue).
• Texturing is simpler for the artist due to UV Maps.
• Trivial to make an object with all over alignment with a mesh. For example wood grain,with prims or sculpts alignment between pieces is impossible and further hampered by overlap. This is much, much easier to do on a mesh. Helps with textured effects such as baked on lighting, shine etc...

• Performance benefits

• Naturally uses texture atlas (UV Map) which can quite easily result in far greater texture efficiency than prims.
• No rendering of hidden surfaces like prims/sculpts. No unused surfaces are built into the model.
• Models can be made using far less polygons than sculpts or prims may LOD to. This does depend on the skill of the artist but an object made out many curved prims of sculpts can easily use far more polygons than the same made out of a mesh.

• Visual benefits

• No seams like prims.
• Totally accurate. Sculpts are approximations, prims have to live with overlap, meshes have perfect alignment.
• Less texture distortion on unusual shapes.

***

Fully retexturing an object is also easy as you just have to swap out the texture map for the mesh rather than applying a new one to every prim and sculpt. A mesh can trivially change from black to silver trim without having to call a texture update in every single prim and sculpt that needs to change.

Ideally we would have bones in meshes which would allow animation, this would be massively powerful and allow things like flaps on aircraft and doors on a car to be smoothly done instead of the incredibly awkward prim animation we have right now.

***

There isn't any reason a collaborative mesh editor couldn't be built into the viewer. It would take a fair amount of work but there is nothing impossible about it. UV unwrapping would probably have to be handled externally.

***

Another issue is that a object made of a couple of dozen prims/sculpts is far, far more work for the server and viewer than a mesh. A mesh might have more shape data than a prim but it is treated as a single entity, unlike an object which is many seperate entities. This is still true even if meshes had a large amount of variable properties like prims currently do.

If only SL could make direct use of the resulting NURBS models (or *.obj meshes) or any other 3D CAD export format then I'd happy to do just that. So would others. So, too, might any number of companies like to demonstrate their for-sale products here, I think. But as for myself, if I must first create the give-away design as right-and-proper 3D model and then re-design it over again out of squashed out sculpties...that ain't gonna happen. It's just not worth the effort.

I'll bet that any company wanting to set up a store for their own products inside SL feels the same way. I'll bet every last one of them designs their products in some kind of mesh and has those models already to hand. But when they look at SL as a potential demonstration platform and realize the work load of having to redesign those same products yet again the idea falls totally flat.

Likewise I had the idea of modeling Tae Kwon Do forms and Jujuitsu techniques for in-world demonstration using Poser 7. In my uninformed enthusiasm I went and bought Poser 7 for just that purpose. Those funds were totally wasted with regard to use in SL. I quickly found that SL doesn't allow articulation of the hands except for certain fixed poses entirely useless for martial arts. So that idea too fell flat.

All I've accomplished inside of SL is to make a few give-away doo-dads...all for use inside SL only. Not a one of my original hopes to accomlish anything truly useful can be realized owing to the intrinsic limitations of there being no custom meshes or animation armatures inside SL. I gave some thought of switching to RealXtend except that it is Windoze-only. Meanwhile I still find SL worthwhile but only in the social context. It's useless for any kind of real work.

Just look at animations, most of the animations out there are the ones that comes with poser, people have just been exporting them to second life.

I'd vote for a solution that would encourage people to build original content, AND to make them build things better that ugly boxy oversized stuff.

Second Life's greatest feature is also it's biggest curse. Since everyone can build whatever they want, most of what's out there looks horribly bad.

-Ilyara

But if I had to pick, I'd pick mesh import over better prims. Oh of course we'd be flooded with ripped models (just look at Facepunch Studio's models forum, tons of models from other games imported into Gmod and Source) but I'd like the ability to model something in Blender, texture it and then dump it into SL without having to twist a bunch of toruses to get something vaguely familiar. Of course, we still have absolutely no good animation tools, ever since Puppeteering was abandoned. (Seriously! We could use object grouping hierarchies to define a skeleton to animate on)

....Even though, if you do decide to improve prims instead of allowing import, can I recommend adding some actual procedural geometry tools? Like some NURBS or being able to define surfaces from curves or things like that. There are many times where I have two torii edges that I've cut so that they're flat, that I need to run another prim through such that the two ends are flush, but I can't torture it enough to get it to fit right. Being able to, say, take those two ends and just click a button to have a new surface built to connect the two along some other curve would be much appreciated.

The bottom line of this argument itself stifles original content creation by attempting to restrict the available options for builders. I don't see the point of that, personally.

"I'd vote for a solution that would encourage people to build original content, AND to make them build things better that ugly boxy oversized stuff."

Two things: 3D mesh support is a universally saught after feature. AND... you could make a lot of beautiful things with a 3D mesh. Beyond that, I don't know exactly how you can encourage someone to create original content, especially when you can go to Xstreet(sl.com) and see hundreds of examples of copyright theft in one form or another, prominently featured on the homepage. But if you're saying I, or anyone else, should be denied the option of importing my own original content from an external application on the supposition that someone else might steal something copyrighted, I think that's really unfair, and I disagree quite strongly. In fact, it just sidetracks from the point.

RealXtend is a good example of what could be done with meshes, as already mentioned. Maybe LL should try to aquire that technology next...

If implemented properly, mesh import could dramatically improve the appearance as well as the performance of Second Life, and that should be the primary consideration.

Seriously? That sounds awesome, where did you hear about this? Or did you mean modrex?

If all the other automobile manufacturers are putting cup holders in their cars, the debate isn't over whether it's worth your designers' time to put it in, nor how people will abuse the cup holders, but how fast you should rip the designers out of their other capacities to keep customers from leaving you behind. Almost everyone wants cup holders - regardless how superfluous - and there is good reason you can't really find a car without one now.

I think this debate is a sign of an aging program and design team, and frightening if you have a vested interest in the future success of Second Life.

But there are some short-term things present in the prim modeling that need perfection before it's dropped to work on something bigger like this, such as (VWR-9203)

However custom meshes I would really think should be the very next step.

(also sorry if it wasn't called "incremental rendering" I really have no clue how it's called and for my bad english too )

If you worry about mesh filters when importing them, just add a vertex or dimension check... things like removing isolated vertices or the like. Resuming, I am quite frankly fed up of the prim system and the way they will always mess up when the LOD is tuned down.

Residents for Polygon Mesh Support

Please join if you're interested in seeing mesh support become a reality sooner than later.

Thanks,
Reed Steamroller

I would much rather have really good mesh import support that is able import a wide range of models from At least Maya 3ds max, lightwave and blender the big ones. We should also be able to Export any old Secound life objects into a 3ds studio/maya/what ever at least the ones we own . I got tons of old secound life prim stuff I would want to bring into a 3ds max program and smooth out what I could not do before.

I mean as it is now the in game tools are good enough for most people and even then if they want to texture they have to use an outside program like photoshop to make textures and even if you are making sounds you have to use an outside program.

Besides If you are building a big item and want even building of the model doing it in world is more of a pain since you have to change you view and avoid other in world stuff blocking you view and people walking into your view.

So I want high end mesh import support in secound life I think it is what I always wanted. in game tools can be imporved later.

We need mesh support if SL is going to steadily improve in performance. Prims are very hard for the renderer to deal with, since there are so many faces that get hidden inside objects when trying to build complex shapes. The renderer still processes the faces whether hidden or not.

OBJ import is probably the best place to start, since it seems it's supported by just about every modeler out there. We shouldn't have to wait months, years for our favorite tool to finally support some obscure format.

The reason being: the current vertices count in simple regular prims is ridicolously high. even a sculpty has alot of vertices than it needs to have (1024 vertices per sculpty), with the basic primitives being between 256 - 1024). Sculpties are ONLY a temporary solution to something of "ability to import mesh". Alot of the software, that is able to do sculpties en mass (such as maya) cost -alot-of-money. Sure you have blender, but there is no proper exporters available (maya has, an exporter made by the OP, with modified versions of it by community members (animation versions, as well as refined versions posted by Leben Sch.).

A typical object in SL may have a ridicolous amount of polygons rendered to get the effect. (64 m radius from a high traffic area may contain more than 500 k Polygons)

The ability to create quality commercial objects in second life en mass using a (high cost) professional program has possibly even encouraged piracy in the sl community, as everyone attempts to get ahead of anyone else. IF however we could simply export an object, and able to apply bump, specular and texture mapping to it (let just say, basic shaders) then we possibly could get the rendering engine to work in a (semi) optimized enviroment).

the feature should NOT replace regular prims, as this would break content, but in my opinion should phase the regular prims out, so that at some point, you do not have to use regular prims anymore.

Ofcourse, if mesh importing would be allowed, i suggest there should be a polygon/vertice limit per uploaded object, as well as overall file size should be optimized, if not converted into a SL rendered format. There is the issue of content security there after. The imported mesh wouldnt need to support animations, as all of these would be done script wise in second life. simply make the parts (that move separately) and then import it and animate it using scripts.

Either allow us to import meshes into second life, or allow us to modify the vertices them selves on a model.

http://www.dominodesigns.info/second_life/blender_scripts.html

I would love to be able to import meshes. Sculpties have their uses. Meshes would be really handy.

I have seen sculpties misused. There are some region builds that make use of them and render is WAY slow. I suspect the same thing can happen with meshes. The vertex limit is probably a must.

" The ability to create quality commercial objects in second life en mass using a (high cost) professional program has possibly even encouraged piracy in the sl community, as everyone attempts to get ahead of anyone else. IF however we could simply export an object, and able to apply bump, specular and texture mapping to it (let just say, basic shaders) then we possibly could get the rendering engine to work in a (semi) optimized enviroment).

the feature should NOT replace regular prims, as this would break content, but in my opinion should phase the regular prims out, so that at some point, you do not have to use regular prims anymore."

Thanks for saying it in such a clear manner - the problem with this whole idea is that it will drive out any amateur content creation. Only the skilled (with outside programs) need apply for content creation.

Second Life got a large part of its start with Philip Linden walking into meetings, opening up a laptop running Second Life, and clicking the create button to get a cube on the ground. "It all starts with a cube." was the motto at that point. It showed that anybody could be a content creator, not just those trained on specialized programs. What we need is not mesh import - if we're going to have meshes, what we need is a new mesh prim type, and tools to edit it in world. Or perhaps the ability to convert existing prims into meshes, and edit them there - but let's stay in-world, and not spend so much time out of the system! It's bad enough that so much clothes and skin creation is done out of the world.

In fact, perhaps Linden Labs could sell a client with mesh importation as a separate product, to larger companies who's primary interest in it would be the importation of their models for design and testing purposes.

IF however we could simply export an object, and able to apply bump, specular and texture mapping to it (let just say, basic shaders) then we possibly could get the rendering engine to work in a (semi) optimized enviroment).

We couldn't optimize the rendering engine any better - most of the non-optimization comes from the fact that the SL Rendering engine has to work on user-generated arbitrary content, so many of the optimizations that game designers use just won't work. To this day, there's a lot of people who never realize that in the earlier episodes of Quake and Doom, you never actually had a fully 3D environment where you could cross over your previous path. SL uses alpha blending in most cases, instead of the easier to handle alpha masking. Things like that means that other approaches to optimization are required.

Personally, I'm waiting for the day when we can rip out the current rendering engine and put in something based on ray-tracing. THEN we'll see something really nice!

We couldn't optimize the rendering engine any better - most of the non-optimization comes from the fact that the SL Rendering engine has to work on user-generated arbitrary content, so many of the optimizations that game designers use just won't work. To this day, there's a lot of people who never realize that in the earlier episodes of Quake and Doom, you never actually had a fully 3D environment where you could cross over your previous path. SL uses alpha blending in most cases, instead of the easier to handle alpha masking. Things like that means that other approaches to optimization are required.

When you are able to import meshes from outside of sl, you can optimize the amount of polygons used in the mesh, and the amount of textures to only a few (instead of per face). Right now as i said in my comment, the regular prims dash out a huge amount of rendered vertices, which adds to the lag of there being multiple amounts of textures per each prim face.

The SL rendering engine it self is a decent engine, as its able to render hundreds of thousands of polygons*, and hundreds of textures* (*dependant on the system) with barely crashing (which it still does pretty often). IF we could somehow edit the vertices, or import meshes, in a manner that the vertice cound is optimized (in second life, most objects have 300-1024 vertices per PRIM) , there will be a lot more detail. However, the fact is that there more vertices there are in the scene, there is more power required to have ray tracing. There are clients existing for SL that enable a limited base raytracing with real time shadows and shaders, but these all have a profound infact. But with the amount of vertices and textures in a scene, the amount of power required to render the scene is enormous, and usually just becomes a lagfest.

I have to stand corrected on my comment on phasing out the regular prims thought. As TaraLi Jie mentioned, she/he is correct that it may cause new content creators to stay clear of creating anything..

However.... I think the regular prims should also be optimized in terms of vertice count or use. (maybe separate meshes for different states, (Solid, Hallow, Cut, Complex(Twisted, Flex) ) per EACH prim)) Ill be making a JIRA Article on this suggestion.... after i check if there is another article on this topic.

Personally, I envision it happening in 3 steps.

1. Basic meshes. Including nothing more than geometry, and UV data. They would act as an alternative for sculpted priims initially. add a new "mesh" asset type in the inventory. dropping one into a prim would simply set it's geometry, like a sculpt mapallowing more to be done with less. I do low poly modelling in Civ, and I can make a human figure, complete with weapons and hair, in under 700 faces. By contrast, a SINGLE sculpted prim in SL is 1024 faces. The inefficiency here is unbelievable. Without having to constantly work around the limits of sculpts, I could make things using far less resources than I do currently. UV mapping would also allow for much more efficient texturing, and reduce texture download sizes. Also, mesh imports could be exact, rather than being limited to this arbitrary 8 bit precision.

Of course, to prevent resource abuse, there would have to be some limits. In the interests of backards compatibility then, it seems logical to allow 1024 faces per mesh, same as sculpted prims currently have. But unlike sculpted prims, we wouldn't HAVE to use every single one, and indeed the vast majority of meshes would use far less.
----------------------------------

2. The second stage, perhaps months after the first, rigging, and custom avatar meshes.
In this stage, seperate the Linden Avatar Mesh, from the Linden Avatar Skeleton. When editing appearance, allow them to be chosen individuallly. Then, allow rigging data for uploaded meshes. This could have several effects.
a. The "attachment" system could be done away with almost entirely. Rather than attaching to specfic points, objects could just "attach to avatar" and the vertex bone weights would tell them how to move. This would allow for example, attachments that are connected, like a backpack with a hose that goes to a flamethrower in your hand. Or a full body suit of armour with no gaps, As long as the vertex limits are stayed within, we could do amazing things.

The other thing this would allow, is that with rigging, we could create new meshes for the avatar itself. Obviously, this would only be of use for mostly humanoid meshes. Things that would move easonably with the linden skeleton. But it would see widespread adoption for furry avatars, robots, demons, and other asorted humanoid-but-not-human designs.

------------------
3. the final stage, complete custom skeletons. Both for avatars, and objets. To begin with, for avatars. user made skeletons would replace the linden skeleton entirely. Since animations should work based on bone names, existing animations would still work with a new skeleton, to a degree. although new bones would obviously not be animated. In such cases, they'd have no animation data, so that other animations could take over. This would allow, for example, an angel avatar with animated flapping wings (via an AO) to still use normal dance poseballs in clubs, and such. Completely new skeletons would also see wide use for completely inhuman avatars. quadrupedal animals, dragons, strange alien beasts. finally, no more would imagination be constrained to humanoid forms. Wonderful things could be done.

the next logical step, following on from that, would be to allow objects to have their own skeleton, and play animations on themselves (via scripts). This would allow for much more realistic obects in general, like rotating gun turrets on a tank that don't have to be a seperate linkset. Or an object that assembles itself into something complex from a box.

Well, anyways. As can be seen, I have big ideas. but I know it's best to start small. So, basic mesh import to begin with, and go from there. I hope it comes soon

Also, isnt .nif like .mdl? a vector/point cloud?
If so then any discussion of skeletons in SL is totally moot.
Even if it does support dynamic realtime skeletons youd need additional bones to do your flamethrower hose. And at least two attachment points to fix them to. Or a buttload of additional skeleton weighting data to specify how much each bone effects every vertex. Or a means of specifying the falloff distance per bone... mostly impractical in a streaming, legacy supporting system like SL.
Also consider that any alteration to the avatar mesh is a radical change, and would have to be in addition to the existing one. Thus introducing a new huge amount of data to everything.. dataservers.. rendering.. problems... steeper hardware requirements, etc etc etc etc

Personally I'd love to see a new av mesh, but in terms of priority I seriously doubt its anywhere near the top of the list.
Significantly higher up the list would be flexible sculpties. As they are more obtainable, and do form a good alpha for what you outlined above.

I want to see LL work on mesh import, specifically as it relates to variations on the already proven image based sculptie format. Not as it relates to import of raw mesh.

I have seen far too many horribly inefficient triangle meshes built by far too many amazingly self confident 3d artists to think this would work in the SL environment.. and I know LL would be falsely accused of incompetence after we started importing some of this complex and attractive, but inherently broken mesh data. Just as they are accused of stupidity by people with 100k+ item inventories, or 600+ script hair.

Sculpties are good. And they need a little more tweaking yet.
Lets wait for the hardware to catch up a little, or some shader magic to get written so we can start flexing and boning them a bit.

SL (outside its core) isnt the latest game engine, built by professionals under hardcore quality control of all its assets... its more like IRC: Slow moving, extensible, community based, legacy supporting. Roll your own. (extend the definition of 'SL' into OpenSim and all the various spinoff technologies)
It's a trivial and completely seminal communications media. Operating with both it's own operational business constraints, as well as those of supporting a huge community of content creators/consumers.It's not some new toy that is released as 'new' each time and forces us to upgrade our hardware to within a 12 month tolerance of its current software release. (not that RTS's do, but many ZOMG-ItsTheNewCrisis' packages do) And it's so much more than an isometric RTS game. (no offence intended)

also;
:]

Example: When HL2 came out in '04, in all of its DX9 glory, it was still backwards compatible with DX7 systems.

If people are making the leap to Blender or Wings already to do sculpties, it is a small step for them to start making arbitrary polygon meshes. I think the hardest point would probably be setting up their own UV texture space, but a lot of software has auto-UV utilities anyway. I doubt there would be much difference in learning how to do either (poly mesh or sculpties) from a newb's view point.

I know, everybody is screaming about how you'll be able to grab stuff from turbo squid or something and upload it. Or rip models out of Doom3 or something. That already happens with textures. The metaverse hasn't self destructed because of it.

I would assume that LL would include some limitation at least to how many poly's a particular object could have, if not allow users to limit the amount of polygons being displayed by these objects in their viewer.

As far as I know, creating your particle effects isn't all that intuitive of a process. It takes time to learn what you can do and how to do it. And, its not like you can't abuse particles either. How often do you think Sand Box Island gets annihilated with a million Ninja Turtles?

(sorry about the update, I'm obsessive over typos. i'll probably find another later and have to tie my hands to my chair in order NOT to fix it.)

To BETLOG Hax:
Well sculpties were a nice temporary alternative, it was cute on the paper and as a concept, but lets face it, it's nowhere near good, it realy can do what it was suposed to do at first, organic shapes such as stones, or fluffy pets... but for the rest it's way too much laborious. It can do nothing realistic, help in noway for buildings, and it has simply way to much limitation YET to hope tomorow it'll change for the best. Now staying logical, having the ability to do realistic things means we can do less realistic things too, what can do the best can do the less, so don't worry about your fluffy pets and your stones, you'll be still able to make them with meshes. You can start to hope they will add boning/skinning into the mesh importer if it's what you're waiting for with sculpties.

TaraLi Jie:
Second Life got a large part of its start with Philip Linden walking into meetings, opening up a laptop running Second Life, and clicking the create button to get a cube on the ground. "It all starts with a cube." was the motto at that point. It showed that anybody could be a content creator, not just those trained on specialized programs. What we need is not mesh import - if we're going to have meshes, what we need is a new mesh prim type, and tools to edit it in world. Or perhaps the ability to convert existing prims into meshes, and edit them there - but let's stay in-world, and not spend so much time out of the system! It's bad enough that so much clothes and skin creation is done out of the world.

Now that's twisted, from your point of view SL should be a "free pshotoshop/3d studio/3d animator/..." when in fact it's already a pain to get decent graphics, and we're in 2010?
The storry of the cube was nice, but I don't think philip wanted just to limit resident creating cubes ? Do you believe in lets say 10 years, resident wll be still playing with cubes and ISL textures? I agree totaly about creating enhanced tools inside SL, but how long it'll take till it cathes up actual technologies?
Just some simple questions that may make you think differently about this issue, is your car working on steam ? Is your TV black and white ? Do you still own a VHS player ? Are your harddrives smaller than 1gig ?

Basic prims are great for basic stuff, but SL has grown out of them. It grew out of them a long time ago, but late is better than never. Sculpties are a stop-gap measure to get the idea into SL, but total Mesh support has to be the ultimate goal of this process. Mesh support (and perhaps a tool for creating, editing and skinning 3D shapes in the viewer, for those of us who can't afford the software), coupled with bigger prim limits, both for physical and nonphysical objects, bigger prims, and larger max link distances would be a great overhaul for the building system, and start SL on the right track to completely revamping it's look for the better.

Now that's twisted, from your point of view SL should be a "free pshotoshop/3d studio/3d animator/..." when in fact it's already a pain to get decent graphics, and we're in 2010?
The storry of the cube was nice, but I don't think philip wanted just to limit resident creating cubes ? Do you believe in lets say 10 years, resident wll be still playing with cubes and ISL textures? I agree totaly about creating enhanced tools inside SL, but how long it'll take till it cathes up actual technologies?

Just some simple questions that may make you think differently about this issue, is your car working on steam ? Is your TV black and white ? Do you still own a VHS player ? Are your harddrives smaller than 1gig ?

The point of the anecdote was not the cube, but what the cube said - simple to start in, easy to collaborate on. No, Philip did not want to limit us to cubes - that's why spheres, pyramids, tori, rings, and other prim types were introduced. That's what my push has been in this entire discussion - trying to keep it in-game, where regular users can actually make use of it, instead of letting it devolve to the state the World Wide Web has gotten, where it has to be all-singing, all-dancing. Frankly, I've seen that world - and we once called it Geocities - now we just call it the Web in general. Every so often you stumble over a webpage proudly proclaiming "Created in vi" or "Created in notepad". You know what? Those pages are actually useable!.

In addition, I don't know what you're calling "decent graphics" in SL. I rather like what we have. It's not 100% photo-realistic - that's quite some time in the future, when someone gets off their ass and moves ray-tracing into a CUDA solution, and scenes get shadows, refraction, and translucency almost for free!

I'll also admit that there will always be the need for external tools to create the most complex projects. But I don't think that need is nearly as big as most people seem to think it is! For that matter, SL is only 6 years old - I think we're years from seeing the best and most amazing results that can come from the simple prims you seem to turn your nose up at.

Never fear - we will get meshes. I just think it's going to further the AOL-ification of SecondLife.

so don't worry about your fluffy pets and your stones, you'll be still able to make them with meshes

LOL, I have no confusion whatsoever about what I can and cannot make with sculpties or mesh, ive been doing the latter since about 1987. (the monkeys were sculpted by a dead friend, and i don't make rocks except for giggles)
I see sculpties as no different, just a format that exists in a system that requires technical constraints to help prevent users doing things that may grossly destabilize the system.
Sculpties, with their known packet sizes, and 100% UV/texturing efficiency are perfect in that context.

• As I understand it; packet size dictates everything in SL.*

http://www.sluniverse.com/pics/Search.aspx?Search=Lindsey%20Warwick%20By%20BETLOG%20Hax

Also, if you hard edge a sculptie in the right way, and then reduce its color depths significantly (by reducing its contrast)... i'm pretty sure you could produce a sculptie that has most of the desirable properties mentioned above. The fact that LOD and the client would decide what to render would reduce the vertexes in this case.

I probably should mention that up close a box prim has 3x3=9 polygons or 3x3x2=18 trianges per side, totalling 18x6=108 triangles.... this is primarily for LOD, flex and texturing afaik.
A sphere appears to have 24x12=288 polygons or 288x2=576 triangles.

A standard female SL mesh i'm looking at in blender right now is ~3900 vertexes (~7700 faces).. but i can't be sure i didn't lose any when i was cleaning it up, so numbers arent intended to be totally definitive.

However I dont think this is about efficiency in as simple a case as people would like to think.
Its more about efficiency with reference to many less obvious factors, Powers of two, packet sizes and a format that inherently imposes the constraints necessary to stop people making geometry thats unhandleable and causes problems...

That model of the girl is awesome. I also realize that you admit building her out of sculpties isn't the most efficient task in the universe (at least on the client side, server/transmission side of things are a different story).

However, in the spirit of this being an open forum for the public, and clearly showing both sides of an argument, allow me to demonstrate what ~7000 polygons does for characters elsewhere.

check this out:

http://bit.ly/h6ZuI

or this:

http://bit.ly/Bi3gG

I mention this only to make sure everybody knows that there is a pretty good reason for many of us to be decrying sculpties as inefficient. Because if you don't tell them, they won't know.

Admittedly, those models are utilizing DX 10 and shader model 4 blah blah blah.... so it isn't just the polygon count that is making them look awesome. Then again, it would be great to see a little of that in SL as well.

Also, without a good/efficient way of storing, and then transferring models such as these to the client, users WILL be waiting around "forever" to see the desired content.
Incorporating standard LOD methods would mean one model really consists of four (or so) separate models of decreasing detail, thus increasing both storage space and transfer time.

UPDATE (Because you knew I couldn't help but to edit my posted message ):

I wasn't going to give Crysis as an example, because its characters are made up of around 10-15k polys. But then I realized "Hey, thats only like 7 more sculpted prims!"

http://bit.ly/ej0Bd

UPDATE 2 (lol):

Clarification: I know, SL isn't DX, its Open GL. I don't want to see SL go the way of DX either (not that i think it ever would).

Betlog Hax:
I'm sorry you took the "fluffy pet and stone" story personaly. But again, I'm not talking with no reasons, it's because I worked with them A LOT that I can say such things, the more I worked with them, the more I saw the limits again, and the more I understood that this technology is realy just made for fluffy pets and stones (it's a caricature, I know many things can be done with sculpties,... depending on what you accepted to sacrifice).

To illustrate that I'm not here just to talk trash about sculpts and prims, here's a build I made only out of "prims and sculpts", maybe it'll make you understand that somehow I know what I'm talking about.:
http://www.flickr.com/photos/39412035@N08/3618132435/in/photostream/

IMO it's a good exemple of the limitations SL has, since I can't texture such a build (over 1k prim), since it's not efficient, I get millions of triangle when in fact not even the quarter of them are usefull, it brings lag... (here I'm realy talking about the prims AND the sculpties)

Anyways, if it's to say good things about the "prim" system, I guess it's definitly a good pedagogic tool, to start "touching" on building. Understanding how tools can work together, understanding that you can put textures, and what not. (well prims are like lego )

I don't think they will remove this from SL, but it pretty much is at the limit of what it can do, now like me, I'm sure you all know that if LL starts a new system to build inside SL, it'll be not that different to what we have right now with prims, it's gonna be again one of their "funky" technologies like sculpties, that we will (us, artists) fight to understand, and in no time see again the limits, and then the same story will just start over again , we'll be again begin for mesh import. So if we can just jump that nonsens mess, and get where it will go at the end without nostalgic thinkiners to stop the machine, I think this is a good thing.

Oken Hax said it all about Sculpties, you can make nearly whatever you want out of them if you "don't need too much realibity", hence my pet and stone story, those are the thing that mostly just don't need precision/sacrifice, and I'm not talking about the hard edged sculpties, you loose directly textures in the 4vertex = 1 vertex process (hard edging), leading always to sacrifices.

Now reed again did point the real thing. Technologies are there to produce verry nice content, and the verry verry first step in order to achieve this is mesh import, he proves it with just some pictures, he even shows a HalfLife2 picture (Alix) and believe it or not , this game is 5 years old... already !

Sl is like our kid, we work on it, think sometimes what we can give him to be better,... And my real siwh is simply to enhance it, to give "my kid" whatever tools he needs to be the best, I do care about secondlife, not in a selfish manner or anything, and god knows if sometime we pick the bad things to make him grow, I'll be the first to say "okay sorry I made a mistake". So again, realy, Just want the best out of SL, don't you all ?

In fact it is "made" with scultpies and prims (inside 3DS max with Prim Composer, I even took screenshots of it inworld after importing it), and making it only out of sculpties is impossible since sculpties can't have hard edges and details at the same time keeping decent lod on small pieces, And for the question "why you want to build it", I just don't get it ??

T M :
Why do we need meshes? Will they be used to make things easier, lighter and better for everyone? Or will it be a wholesale download of turbosquid? :

Yes to both questions. Now I don't see you complaining about "botters" and "skin ripers", and.... Don't oppose those 2 ideas. You know good things comes with their opposites, and none will "win" over the other one, that's how world goes since the begining, and unfortunately, you and me can't rewrite how things are working.

T M:
Qarl was one of the first to say that we had turned sculpties into something it was never intended, a mesh system.

A mesh system that fails if it tries to act like a mesh system, it's all about sacrifice/limitations/Lod controls/... Man we're not talking about the same thing. You talk about things that work and there's is no doubt they work, BUT I'm talking about things that can be better.

T M:
I personally now believe anyone who says they can't do something in world with prims and sculpties, doesn't want to invest the time.

You realy have no idea, I'm far to be a person who's gonna spend a minute on something then give up, don't believe I haven't spent at least a week on mastering sculpties because yes, they have no secrets for me. That's even the reason why I can say "A better system is definitly required". I mean I know all the limitationS they have, and what sacrificeS has to be made for workaroundS, and THIS is exactly the point on having a true mesh system.

T M:
So the question is still why? Isit because you don't want to spend the extra time texturing individually?

How do you want something that has over 1k prim to be textured, each prim individualy ? (each prims having about 6 faces, 6000 textures to upload? workaround with faked ambient occlusions and stuff?) Optimising it, .... Keeping it efficient, lag free, ... I can go on for hours. Now if you talk again about workarounds... you know my point of view on this.

T M:
Also few have legal access to those $1000 programs.

Come on, you're not the first to talk about this issue, I can not say it enough, blender/gimp/... and many others DO COMPETE with verry expensive softwares ! (google for movies made with blender, verry impressive)

T M:
Is it because you lust after crysis and want SL to look like that? Or is it because you really want to help the people to enjoy their time here and help continue their ability to add to our special world

Again, both are true, now don't oppose them like they're linked or not..

I see you already used twice that "political" technic to compare things that have or nothing to do together, or are simply the same thing. This is far to be a constructive conversation if you go the politician way. When you buy a car you know it can take you to some place, and someday MAYBE make out of you a murderer, but you're not like "should I buy a car to go there OR should I buy a car to be a murderer", that makes non sens, you know already the risks you're taking when you go buy a car/jump on a plane/... there's always risks at some point.

But for that last comparaison you made, why would a "SL looking as good as crysis" impossible with "helping people and let them continue their ability to add to their special world" ? You realy should explain this to me.

T M:
If the Virtual games worlds represented countries on earth we would be the US. There is China. as we move forward lets remember that no company has been willing to give the amount of freedom we have been given, not even Blue Mars. (open grid doesn't count because it basically is SL on different servers. It was born of sl)

I will just quote TaraLie Jie:
Second Life got a large part of its start with Philip Linden walking into meetings, opening up a laptop running Second Life, and clicking the create button to get a cube on the ground. "It all starts with a cube." was the motto at that point. It showed that anybody could be a content creator, not just those trained on specialized programs.

This is in fact the whole concept of SL that TaraLie Jie is talking about, so saying "lets remember that no company has been willing to give the amount of freedom we have been given" it's completely void as it is the main SL concept.
Now you come again with your "weird" comparaison/opposing habit, taking 2 things that are totaly not linked, instead of comparing SL to blue mars, compare SL to RealXTend. Both SL and RealXtend fights in the same weight category... Blue mars is in another one (another concept)

is impossible since sculpties can't have hard edges and details at the same time keeping decent lod on small pieces

Yes, they can.

Each letter is a single prim. Hard edged. Reasonably LOD resistant that you wouldnt really notice it snap levels at distance.
http://www.sluniverse.com/pics/pic.aspx?ID=378665

Small, and approaching the limits of sculptie resolution, note the wiggly edges in some places. The grid behind object this shows 1cm (0.01m)in 0.5mm (0.0005m) increments. Each letter is one, or occasionally two prims.
http://www.sluniverse.com/pics/pic.aspx?ID=378673

Similar size objects, very LOD resistant. One prim each
http://www.sluniverse.com/pics/pic.aspx?ID=378674

Also tiny. Red box is 1cm (0.01m)
http://www.sluniverse.com/pics/pic.aspx?ID=378672

Variation on previous objects scaled up for clarity
http://www.sluniverse.com/pics/pic.aspx?ID=378671

These were made before oblong/ratio'd sculpties, so could easily have better curvature.
http://www.sluniverse.com/pics/pic.aspx?ID=378666
http://www.sluniverse.com/pics/pic.aspx?ID=378668
detail (theres some regular prims in the underlying handle)
http://www.sluniverse.com/pics/pic.aspx?ID=378667

Some high-texturespace-efficiency single prim guns by Gutterblood Spoonhammer.
Holster is one prim, gun is another prim, the only additional transparent box prims are to allow various particle effects to happen simultaneously.
http://www.sluniverse.com/pics/pic.aspx?ID=378670
http://www.sluniverse.com/pics/pic.aspx?ID=378669
Gutterbloods style is minimal texture size with maximum impact, essential in any professional development studio, and conducive to fast FPS in a multiplayer environment, so before anyone criticises that they look a bit pixely we should all should learn a few things from this inherent discipline.
It's easy to make stuff look better. But making it look awesome with minimal resources is what this whole thread is about.

I am absolutely positive that most people want mesh because they simply dont comprehend what can really be done with sculpties.
I am not against mesh as such. But sculpties are a very adequate solution. VERY very adequate.

[edit] adding another annotated example http://www.sluniverse.com/pics/pic.aspx?ID=378677
Also; Meh Hypatia. I'll be only too happy to make efficient sexy mesh goodness when its available... and rant about how badly the usual suspects are doing them... But I'd prefer the basic interface to work properly VWR-7972 before we start trying to do serious geometry. And sculpties are very decent given the current constraints of media such as this.

The main issue is that people will upload every high res thing, and the fact that some people are tools (the dumb variety) needs to be dealt with, or at least planned for.

I am absolutely positive you simply don't comprehend what can be done with meshes.

Custom UV - prims do not support custom uv maps - this saves on texture downloads. Imagine that you baked a texture for a 30k polygon huge house model on one 1024 texture atlas. Same thing on a prim based house would be a significant number more for textures, resulting in more hits to the asset servers and more download for textures as they can't be made to be as easily efficient.

Rigging - save on texture downloads - sculpties can only be effectively animated with series of sculpt textures. This is wasteful. Only way to avoid lag is to have ALL the sculpts rendered and then made invisible and use texture switching. Otherwise you get the happy ball effect, even sometimes when you put the sculpt textures on a transparent prim. This is horrifically wasteful.

Low-rez mesh - less stress on your graphics card. No discarding of "hidden facets" when they simply aren't there to start with. They could be set to LOD in a much more sensible manner as well.

And lets get to another major point. Real life businesses who want to import their data for visualisation, cannot easily do so in Second Life. It's why I have moved on over to messing with other engines. Which can.

However, I am a supporter of having a wide variety of tools so that one can choose the tool that best fits the job. Sculpties have their pluses, but they're not meant for architecture, and they are not ideal at animation. They were meant for simple organic models, to reduce the number of total prims used in building, most especially wearable attachments such as shoes and hair. Those who promote them for uses they were not originally devised for, do not realise how inefficient they are compared to other options.

It's like working with a hammer with a screw when we'd better use a screwdriver.

Well first thing first,

http://www.sluniverse.com/pics/pic.aspx?ID=378665
You can easyly notice the poor precision of the sculpt file , this would not happen with meshes
(I KNOW it's due to the 255*255*255 position limit, don't get me wrong, I'm not saying those are ugly sculpts, in fact those are verry nice sculpts)

http://www.sluniverse.com/pics/pic.aspx?ID=378673
The sculpties precision limits are far reached and are verry obvious. (again, those are verry nice sculpts!!, but the sculpt system just fails, and this picture is a real proof)

http://www.sluniverse.com/pics/pic.aspx?ID=378674
It seems like you mastered the limitations too, but honestly "snailing" was fun for you ?I do snailing for doors/windows and such and it's just plain user-unfriendly (real PITA), is that the way to get "builds in sl" affordable for everyone ? I don't even believe regular SL user can even imagine the workaround we use for such works (by the way, your chains are verry nicely built)

http://www.sluniverse.com/pics/pic.aspx?ID=378672
http://www.sluniverse.com/pics/pic.aspx?ID=378671
Verry nice diamonds, nothing else to say, so yeah, definitly proving that sculpties are best suited for stones and fluffy pets. (just kidding, don't take anything bad)

I realy do like your sword jobs and believe you did a great work with what you had.
However the Gutterblood spoonhammer ,hmm hmm, I hope you realy didn't take it as an exemple of what can be done with only 1 prim. That sculpt must pop in lod levels like crazy and saying :

Gutterbloods style is minimal texture size with maximum impact, essential in any professional development studio, and conducive to fast FPS in a multiplayer environment

Is totaly untrue. 10 years ago I remember I had my 3K limits for game caractere modeling, but that was 10 years ago. And controled lod + the power of extrusion, the ability to jump to concave and convex mesh are just some of the many things that are simply impossible (or verry tricky) with sculpties. (f.e. snailing like in yoru chain, like in doors,...)

I am absolutely positive that most people want mesh because they simply dont comprehend what can really be done with sculpties.

This is again plain sensless since the purpose here is to get resident having the "verry over mega complexely difficult" 3D importing stuff, or the verry easy sweet soft cuddly sculpties that only "you" seems to master (If I listen to what you're syaing). Again .. I do master sculpties, this is not about knowing how to handle them, this is about "imagining A / building A / getting A inworld"
I can only admit that sculpties is a form of art, and I'm glad you like it, personaly it pisses me off more often than satisfy me AND this is because I ASK THEM things they simply cant do, which means i reached my limits with them, and trust me, I'm not alone, YOU DID TOO.

Bet probably likes sculpts, just because he's mastered them. He's finally figured out how to make them, and likes being king of the hill, but I'd also like a shot at 3D shapes in SL, but from what I'm hearing, it's beyond my skill level to produce anything decent. That's why I would like Mesh support. I would also like an editor for them in the viewer, like LL promised with sculpties, but never delivered.

As a matter of fact, perhaps if they made the editing tool good enough, they could completely remove mesh importing, and make mesh creation totally in SL, like with prims, removing any argument that it could be used for copyright infringement,. But, this is an idea. Considering their reliability with promised features of this type, they'll probably have to make meshes importable, because we'd never see the editting tool.

Inworld tools for modelling mesh isn't really the answer. Better inworld tools are needed for a lot of things, but not for things where offlline tools are superior. What we need are better access to offline building tools when these make sense, and a better workflow for artists overall. A major problem I have is that I have to make things especially for Second Life, and I cant easily reuse my work for other purposes.

I would stick to improving inworld tools in the areas where one needs to use parametrics or wish to create dynamic content - I was severely disappointed when inworld posing for avatars was dropped for example. People would have been so much more expressive if they could have posed themselves inworld.

I just hope Qarl's superiors agree with this.

I do however have some knowledge of the technical discipline it takes to workably fill a render scene with geometry. Because people who are new to this lack the aforementioned discipline the geometry format itself needs to constrain them to reasonable limits.
Because like Hypatia said: if sculpties are too hard, raw mesh will be totally out of the question for the people who will no doubt start producing it anyway. Often on a commercial scale. Thus subjecting us all to its potential horrors. And invariably labeling LL incompetent or something ironic because of our inherent ability to blindly exceed necessary boundaries.

I'd like mesh too. But sculpties are pretty damn good, and I'd prefer the basic interface to work properly first.
Fixing the minor interface inadequacies is a small step, and one that should be taken before a significant leap like mesh import.

...I guess i think they should happen together, existing tools fixes being so comparatively minor, and mesh import totally requiring a decent toolset.
My postings here just reached a full circle.

Also; blah blah blah, blah blah blahdy blah blah

If you maintain the correct vertex counts, and understand the 256 increments of resolution in each of 3 directions (RGB / XYZ) then what you make is 99% what you get.
Very commonly people seem to assume that a more expensive program, using more vertexes and the inevitable handful of file conversions will ensure better results. This is why they fail.

Understand the format, build to its specifications, and its very simple to get what you want within reasonable tolerances.

Let's go in numbers... uhhhhmmmm... a sculpt is around 2000 polygons, more or less... and there are stuff which can be done in less than 1000. I have done some modelling for Half-Life, Jedi Knight series and Unreal 1, semi-old school'd games (we can say it like that) which teaches you to be an efficiency whore, stuff like guns, for example, or even characters, which were done in 1300 polygons... 1400... 1000... but with sculpts I am restricted to use those 2k. FORCED, and sometimes more to get the same result, with a poor resolution and horrible texture stretching due to the lack of UV maps.

Why? Sometimes I like to hard-edge, but I don't have something as basic as "smoothing groups", which requires twice the polycount to hard-edge the necessary faces, resulting in an horrible waste of polygons.

Another thing which drives me sick is that there is a polygon waste when splitting the sculpt in several pieces, which can be avoided with meshes.

And of course, the lack of ACCURACY, and wrinkles, which forces you to REDO something several times because you have to ADAPT yourself to the damn system.

Do you still dare to promote sculpts over actual 3D geometry? I think I understand sculpts pretty well, I even use Photoshop sometimes to achieve certain shapes, and other times 3D modelling programs. So I understand 3D as well. I think I can say that sculpts are good, yes, but very inefficient. You're not the only person in Second Life that knows how to make a sculpt, so you should reconsider before going "you fail at sculpts", because that amount of arrogance can be considered flaming and not constructive for this JIRA post.

http://www.youtube.com/watch?v=sxoUwdJQbrw&feature=related

Well no idea if it's "true" at all, but well ... looks like done with many meshes

P.S. @Johanick Hutchinson: I had a look at that video, and ran the accompanying description through a translator. It's actually a copy of an article from another site. From what I've gathered, this is a new 3D rendering engine, complete with a plugin that allows, easily, interactive HTML on prims, that's being released for use in MMOs and virtual worlds.

I've also gathered that LL is apparently paying special attention to it's progress, and may be interested in purchasing it for intigration into SL when released. I don't imagine they'd be particularly verbal on this issue either way, but if you ask some Lindens about it, one of them might give you some more information.

The only thing that I like of sculpties over meshes is that copying it is really hard. Either requires the nasty CB or being able to grab the sculpt texture, then you're doomed, but otherwise, it will be hard to copy and easier for content creators to prove something was created by themselves (since sculpts are restricted and requires special filled sheet UVs, etc)

I wish there was a way to get meshes with that advantage sculpts have... then it would be perfect, but I doubt there would be a way. Is the only "against" I have with meshes...

Now that I've learned how to make sculpties with Domino Marama's blender scripts, I don't feel that sculpties are terribly limiting, except where surface textures are concerned. I still don't quite understand how to combine multiple textures in blender to create one texture map (in spite of the wonderful machinimatrix tutorials on the web).

I lost my momentum, while learning to texture sculpties, because I found it so disheartening that textures were difficult to apply and control.

Except for textures, I'm impressed by the sculpty system.

http://kanae.net/secondlife/

As for taking the building out competely, that would be really bad. I would like a building tool kit sort of like what Dungeon siege had when they first started or like the sims has. (mind you neither are easy to work with. DS's GAS system was difficult because you had to compile the image into their container for it to even be viewable).

Read more about it here:

http://bit.ly/17a5X2

As stated by T Linden, this still isn't going to happen without the momentum of community support. Please comment everywhere you can in favor of mesh support.

It would be a great thing to be able to import a 3d model into Secondlife without having to fuss around with buggy conversion tools and expensive 3d Applications that can cost hundreds if not thousands of dollars. A user may also be limited by their operating system(s) and hardware platforms not being able to run said software * cough* Vista * cough*. I have spent Numerous hours learning a new and many times unproven said to be able to export Sculpt maps or a readable format such as OBJ. This is a very frusterating problem that would make a seasoned buddist monk flail his arms around in a full blown imbolism. and even when you can generate a very good sculpt map you still have to hold your guts and pray that the conversion from TGA, PNG, BMP, etc doesn't go to poo when it gets converted to JPEG2000. We all know those nightmares. It can also take many sculpt maps to produce complex detailed organic shapes compared to a single detailed mesh of only a few thousand polygons. Its true Secondlife's current Building system could use an overhaul, However I think the content creators, the bread and butter of Secondlife's growing economy could benefit by direct import of 3d models, and I speak for any content creator as well as myself that LL should focus on a 3d import tool for developers that could do one or both of the following : import mesh and convert to sculpt map. and/or Import mesh as is.

Every one has my vote for the focus on direct 3d import not export or 3d models.

Remember that making the viewer caused copybot. This will not create a copybot, because anyone can just upload anything from the net and claim it's his or her creation. The end is near!

Mesh support will allow pretty much any 3d software to be used, even the free ones.

Short-sightedness or fear of some possible abuse is not a good reason to limit creativity.

With Meshes the 32 prim limit for physical creations should no longer be a major hurdle to creation. It will free SL from the blockish feel it has now to a true 3d environment.

All hail the meshes!

Samantha Glume

http://blender.org/

But I'm still against this having priority over improving in-world tools, things like VWR-13344 would allow for huge creativity freedom using in-world tools, and it would allow for "it all starts with a cube" to remain a good description of how most of SL is created. Creating things in-world instead of using programs feels much more natural for me, it would also increase the time people stay logged and likely reduce the time required to create stuff (no need to switch back and forth between creating and testing programs). And with the tools being in-world there would be less complaints about the creation certain types of contents being only avaiable to a certain group of people.

Pro's for mesh import:

1) Unless the improved building tools allow you to essentially have all the options of Maya or 3DSM or some similar program "in Secondlife" the improved building tools will be, by definition inferior to Second Life's. I don't think such capacity is going to happen.

2) If LL doesn't introduce mesh support, some other metaverse will. There's a few other "metaverses" being worked on or in planning as we speak. If LL holds itself back, someone else will introduce mesh support. The most talented content creators will certainly be tempted to move to the new metaverse, which is a net loss for SL. I think mesh importation is critical for SL to remain a viable and competitive platform for creativity.

3) A lot of people discuss how the introduction of meshes cuts out "the little guy." I'm not unsympathetic to their point of view, but let's face it, outside programs (often costly) have been used and are virtually considered required for years now in other areas of SL custom content. People regularly use Photoshop, Draw, and similar programs if they choose to exercise their creativity and make textures - it's been this way for years. Motion-Capture animations (requiring expensive mo-cap equipment) now dominate the market. If you want professional sounds you've had to have your own recording and mixing equipment. Nobody is screaming bloody murder about the "requirements" of these other programs. Even in building, sculpties have long since pushed out the "little guy" in high-end content, especially avatar attachments (such as hair or shoes). In all of these cases, there exist cheap or free alternatives that might not have the functionality of the more expensive programs and people do use them (like the GIMP for texture-making).

Con for mesh import:

1) Copyright Issues with people ripping meshes off from other sources (such as games) and uploading them.

http://changingworldsbuildingdreams.com/linden-lab-offically-announces-mesh-support-in-second-life-at-slcc/comment-page-1#comment-5720

People freaking about whats going to get them knocked of the competition need to skill up.

People worried about the ensuing noob lagfest (like me.. a bit) will need to hardware up. (I am reminded why my PC setup usually ends up costing 3-4k AU$... and I'm ok with that given what, and for how long, I do with it, the money I theoretically make with it, and the plethora of awesome it brings.

People worried about copyright issues have a good point. But ultimately it's up to the holder to patrol and protect their I.P. ...In other discussion on this I always say we need a google-like mechanism to trivialize doing that.