We have discussed options internally and believe we should iterate on the Release Candidate by taking your suggestion in this version, 1.20.
Although we introduced Search>Places as a web-based search as of 1.20 RC7, we will return it to the previous state (with "Sort by alphabetical" or "Sort by traffic" options). Look for this change in RC9 or RC10.

The Search project does aim to transform more search functionality to the web-based interface (like the "All" tab). This included the switch from "Popular Places" to "Showcase" (as earlier pre-announced
at http://blog.secondlife.com/2008/04/28/second-life-showcase-popular-places-and-the-future-of-traffic/

The Search project team continues to discuss the future of the traffic metric with the Resident community (and is planning to provide better metrics to Landowners etc); however in the meantime Search>Places should remain similar as before, with traffic sorting that Residents are accustomed to.

Apparently, the new search is broken for a few people; i.e. it won't scroll. This is the only point that Kitty raised that has any merit at all, and it does need to be fixed urgently - even by making the window resizable. But that is not specific to the Places tab.

The Places tab needs to be either changed to the new search, according to the plan, or removed altogether. You have no doubt noticed the continual outcry from the residents about getting rid of traffic bots and traffic camping. The Places (traffic) search is totally gamed by the bots, and is useless as a search system. It benefits only those who, like me (yes, me), use traffic bots and similar, and totally fails regular SL residents. You were finally doing the right thing for the population, by changing it to the much better search system. DON'T allow the false arguments of one resident, plus the votes of a few friends, to cause you to do the wrong thing for the population.

For every point that Kitty made (except the non-scrolling that affects a few people), there is one overwhelming answer:-

"The quality of results is light years ahead of anything else, and it's easy to use!"

You've changed your mind once - please change it again to the right way.

When the Lindens put this HTML search/places into the Release Candidate, they didn't just move to HTML, they also apparently have removed the sort function from search/places to be purely traffic. So contrary to what many fans had insisted, and contrary to what Lindens implied, they did take out traffic as a pure sort function, and substituted it with God-knows-what – a secret amalgam of links, picks – and yes, traffic, but with an unclear means for capturing (snapshots? random assessments?).

So I view this as a stealth political move – and I would like the Lindens to take responsibility for it. Many people have indicated to the Lindens that they DO NOT want traffic removed from search sorting – and that SEARCH/PLACES is PARTICULARLY important as the engine of the economy. While much agida is expressed about the soon-to-be-defunct "Popular Places" as camped/gamed, and while Search/All also contains visible camped/gamed results, Search/Places, especially when used with a rational key word set, is the least gamed and the best display we currently have for MERITED NON-GAMED traffic. Most people skip past any obviously gamed/camped first returns and read the rest of the list for MERITED traffic giving important information about a site's real popularity, desirable content, and customer service.

If the Lindens plan to remove traffic sorting from Search/Places completely, or if they intend to dilute for reasons known only to them, they must be very up front in doing so. They can't pretend that it's a bug, an HTML move, or anything of the sort. In fact, ideally, if they really absolutely insist on moving to a diluted/mixed sorting device for search/places on HTML web pages, then they should also retain the legacy search/places sorted by traffic. And then they'll see from observation what is really used by users, and what really leads to sales – it will be easy to check.

I don't use alts to take part in public discussions or JIRA proposals or group chat – my alts merely hold land groups or perform content/theme-specific chores. And the real tiny minority here is not me or my tenants or my neighbours or my fellow business people in SL, many of whom are for keeping traffic, but the tiny percentage of geeks and devs who have a visceral aversion to traffic for reasons I can only speculate have to do with some fear of losing their power over the software.

There isn't any "research" showing any "99 percent" or even raw majority wanting to get rid of MERITED TRAFFIC. The straw polling done on this merely shows that some vocal people want to get rid of bots and camping and gamed traffic results that ensure. That's different.

That the Lindens, who contain in their midst some of this tiny percentage of geeks with a violent allergy to traffic, would still come out ruling in favour of keeping the sort of search/places option by traffic lets us know that some wiser, adult, business-oriented Lindens remain at the Lab and have some basic common sense about how the economy works, which is not by ideology, but by practicality. And thank God for it, because we can then rationally go on paying tier to them.

This isn't just about some shallow "usability" issue of how the screen looks. This is about the actual viability of the inworld economy and inworld businesses which already face daunting challenges from poor grid stability, difficulties in advertising, a land glut, and Linden competition now in the land development business. This impacts the content economy as well, as content producers need land to display content on.

As I have constantly maintained, SEARCH PLUS PLACES PLUS TRAFFIC = SALES. I constantly poll customers, and tenants who run stores and other businesses also ask people how they came to make a purchase, and the answer is overwhelmingly clear: by search/places. Yes, also by classifieds. But NO, to a far less extent by "word of mouth".

That a few oldbies or weapons sellers continue to make sales by "word of mouth" based on longevity, visibility on the Linden splash screens or whatever can't countermand the reality of the economy: that most purchases happen from search. That's why traffic matters.

This constant doublespeak about "improving metrics for landowners" is entirely unpersuasive. I don't see a single metric the Lindens could add that wouldn't dilute and even cripple the simple, elegant function of search/places/traffic sorts for sales.

Phil, I marvel at how you can impose your own highly subjective and selective concept of "usability" on the general republic. Remove "places" altogether? Huh?! I'm well aware that the most common geek/coder pattern for usage of "search" is to use "search/all" in the belief that this yields the biggest "information dump".

But the average person doesn't need a huge info dump even if the dump comes from more cleverly applied key words. They don't need every single avatar with a weird SL last name like the key word, every single group with that key word in it, and even things like transcripts of townhalls.

Most people – unlike the geeks who prevail here with their very, very obdurate subjective opinions – have no trouble using the Second Life search like ebay.com or amazon.com

Most people expect to use SL not like Google, which is vast and performs huge information dumps sometimes irrelevant to the task at hand, but to use it more like a community of intent, as on amazon.com or ebay.com

Both those services USE TABS TO SORT. So you can pick something like "toys" or "books" and refine your search, and that's how real people do it in the real world.

If the Lindens ever took the tabs off the search here, I'd close the window and walk away permanently, because I know that this company founded by the ebay guy Pierre Omidyar and the amazon guy Jeff Bezos would have lost its mind and become untethered to what those two guys know about Internet sales: tabs, tabs, tabs.

The new Search All is in my opinion a very good tool, and I happen to almost always find what I want. As long as the owner of a parcel optimizes for Search All, they are bound to be found. Whether it is a shop or a nice place to be, I am very able to find it in Search All.

So my proposal would be to either get rid of Places Search at all, or to let it have the same format as Search All (maybe more results per page though). That way every parcel owner can optimize the parameters to appear in Search, without having to use bots and/or camping chairs to get a good listing.

Every Search system can be influenced, to the better or the worse, but the new way Search works, is at the moment the best one available. So reverting RC 1.20 back to the old places search, is a step back. Compare it to reverting back to Windows 95, because you do not like some parts of the Windows Explorer in XP.

But even more concerning: Will this course of action actually displace the new GSA backend to Places search and revert to the old, overwhelmed DB? Something has to alleviate that load before the current partial band-aid of H/W comes undone, or there will be no need for Search of any kind: there will remain nothing to search for and no residents to search. (Can we have forgotten the April concurrency graph so soon?)

(This is about usability, ranking is an entirely different - fixable - problem so the quality of the search results is not relevant here)

Anyone who wants to use the new search all or search all with the places dropdown selected is perfectly welcome too. And in fact, I do this 95% of the time myself. However, there are occasions when I want to quickly scan a large number of results to narrow down my search without the click, wait, back, click, wait, back. Also, as per my note from the cross-referenced duplicate issue 7326, the ability to view on the map and create a SLURL from search without having to visit the location is incredibly useful.

For the record, I have never to the best of my knowledge had any interaction with either the issue reporter nor any of the other voters.

You wrote:
"When the Lindens put this HTML search/places into the Release Candidate, they didn't just move to HTML, they also apparently have removed the sort function from search/places to be purely traffic. So contrary to what many fans had insisted, and contrary to what Lindens implied, they did take out traffic as a pure sort function, and substituted it with God-knows-what – a secret amalgam of links, picks – and yes, traffic, but with an unclear means for capturing (snapshots? random assessments?)."

You are mistaken. Nobody thought that LL had merely changed the Places tab to an HTML format, but kept the traffic rankins, and nobody thought that the new results in the Places tab are anything other than the results from the All>Places search. Nobody deceived anybody. Everyone, except perhaps a few of you, knew exactly what had changed and how it had changed.

Your idea of "a secret amalgam of links, picks – and yes, traffic, but with an unclear means for capturing" is nothing other that a Google ranking algorithm - nothing mysterious about it. It's what makes the difference between excellent results and useless traffic-only results. It's the reason why the change is essential.

My business does not use bots and our number is pretty high for the category we have chosen. We also depend on the places listing because most of our visitors use that to QUICKLY locate social venues that have people in it, not the dead place that paid a ton of people to add them to their profile picks. The new search is misleading and clunky if you are looking to see if a place is currently populated.

That's really the whole JIRA proposal summed up into one line.

This JIRA was meant to be about the presentation layer and how easy/how much time it takes to accomplish the same tasks in the two very different systems and how prone either is to failure. Where the results come from is totally irrelevant for that and is a completely separate issue. I just personally happen to like the two pane navigation way of navigating the results of Search / Places and do use the category search and "Show on Map" quite often and merely pointed out what I think are flaws in a Google-like presentation for the Places results vs the old layout.

Search / All is fine in a Google-like result set in my opinion, but I don't see how it would hurt anyone to keep/revert to the old Places layout regardless of where the results come from. It's all a matter of preference and this JIRA was a feature suggestion to retain the layout of the old Search / Places.

The only reason "non-HTML" was part of it is because I personally don't think the web browser control is mature enough to be the one and only option. It's not just currently about searching, but about even being able to click the results at all. I don't see any reason why the result set can't be based on GSA but not presented in HTML form.

If all you meant was: "We do see some problems with the new layout compared to the old layout for Search / Places as well (not necessarilly the ones you see, or in the same way) and we decided to - temporarily - revert back the old way while we work out a way to present the new search results in a way that is more consistent with the old layout" then the decision to revert back completely (traffic & all) would really make a whole lot more sense.

I really do miss traffic metric. Camping problems aside, right now, I can't search for "sandbox" and visit sandboxes with lowest traffic - which was my favourite search pattern.

I too hate new Google-like layout, but I guess that there are limits how Google search appliance can be customized. Some of this feeling is, granted, connected with unpolished in-client web browser.

On the other hand, I like the new search, and how it mostly returns more relevant results, and I like that there are many more factors other than traffic that influence ranking. I would hate to return back to old places search... tweak the new places some more please.

Usability:
Kitty hasn't discussed the searche's usability at all. She's said that her objections are about usability, but they are not, so please don't imagine that it's anything to do with usability.

What Ramzi said about the new layout seems clear enough to me. It's going back to being exactly like it is in the current standard viewer but with the option to list places either alphabetically, or by traffic. Wow! Haven't we done a good job on this one, folks!

Cristalle:
A vote is a vote. It's too late now for Kitty to clarify what she meant. You've voted for the way that Ramzi & Co have understood it.

The biggest problems with the new Places Search for me are in order of importance...

The lack of the Show on Map button
The lack of the Search by Parcel Category function.
10 results per page is far too few and the layout is messy
The search dialog window is not resizable so can't even view the feeble 10 results without having to scroll.

I'm in two minds on the traffic scores. Whilst abuse is rampant, it was pretty easy to simply ignore the astronomical scores and try go directly to the more sensible ones.

It works to help MOST people make purchases and visit desirable venues and thus works to make MOST sales.

You are in a very vocal and visible minority of users on the forums, with the perspective, typical of geeks, that traffic is "broken" because it doesn't work (so you imagine) like Google, or "do what you want" or "is gamed" by economic activity you don't like, or find culturally low, or whatever.

But it works. Once you take out Popular Places, removing in part incentive to get very high traffic achievable only by bots and camping, and once you merely scroll past the first few gamed slots, or use key words more specifically with the PLACES tab (instead of SEARCH ALL as you may be doing), your sorting improves enormously, camping/bots do not soil the eye, and you do what most people in SL do: you shop, you buy.

Nothing broken about that: it's the engine of the economy.

Confusing SEARCH/PLACEs with a new and non-transparent search sorter without traffic doesn't improve the ability to find what's good. The strange combination of picks, incoming links/landmarks, traffic doesn't provide any coherent ordering. Traffic is simple and effective, and on MOST of Second Life's more than 20,000 sims, is NOT gamed and camped.

Try to get over the rough aesthetics you experience with a tiny minority of parcels from cultures and economic styles you don't ilke, and see that the issue for most people is finding what other people also found good – very simple. Picks just doesn't tell the story, as it is not widely enough used – and also gameable. BTW, I was the first to suggest they use Picks instead of their trafficked Popular Places, and they didn't, while recognizing its value.

Removing tabs is insanity; sites that wish to make SALES and not score IDEOLOGICAL POINTS about formal, geeking programming symmetries understand that amazon.com and ebay.com and many such sites with sorts by tab are what work to help people buy and sell. And that's how it has to be in SL.

Once again, I can only attempt to appeal to common sense here. What on earth is "hopelessly flawed" about an honest metric of avatar visits, not camped, not gamed? What on earth could be construed as "broken" about that?

If camping bothers you – get the Lindens to deprecate the Lazy Lindens script and other camping scripts. If bots bother you, get the Lindens to regulate bots. Don't destroy the economy to rid yourselves of a few unaesthetic features of a free market.

Whilst I'm not against the current traffic system, it does skew the reporting to the heavily commercial sims and therefore is much less of a benefit to the small independent music club, RP sim or any other interesting sim that has no financial incentives other than to pay next months tier fees. I help run a small club and amongst us there is little appetite to fill what space we have with s**t to sell, places to camp and relentlessly go for any old AV.

Runing places in SL is not all about commerce you know Prokofy.

However, I can't think of a decent way to score 'coolness' so in the absence of that I would prefer to choose from a range of traffic metrics (eg unique visitors or current visitors) and a lot more categories to sort by. I would also like some way to identify camping spots and especially camping bots and have some way of factoring out those results if I choose (maybe a tall order there).

On a slightly different tack, what I'd really like to know is how much of the Google search engine is implemented. Can I still use all those acane advanced search terms of the real engine in the SL version? If not, which ones can I use? Are there any resources anywhere? For example will OR work, can I search by title alone, can I find the most recent, will inurl: work to find sim locations, etc, etc?

Also as a normal SL user I find the new searches less useful than the old ones. I don't find my old places anymore. Also if you want to warm people up for this search you also have to publish the exact criteria to get a rank.

I think traffic is still the best of the bad solutions. Also the new search funnily does not support standard search criteria we would have hoped for. (At least my tests were negative). If one wants to ad the keywords "Blue Sky" and "Blue heaven" to ones site the non repeating form in "About Land" like "Blue Sky Heaven" still does not show in a search for "Blue Heaven". One needs to have a repetition in the "About Land" in the form of "Blue Sky Blue Heaven".

I have two proposals to help people. Someone who can pay campers has also the products that have the quality of products that sell well. So it is ok for business. But I hear the voice of the one that quotes that "non profit" places lose. That is a big draw back (which btw is also not fixed in the new search. I don't understand priorization at all there. I personally go back to the normal client because I simply think I get mainly "crap" places and/or the most expensive classifieds (hint hint)).So proposal one is easier to implement: "Give another criteria additionally to the Category, "non profit" or "service" or "Shop". This helps everyone. If I look for a nice sim to explore I don't want business hits, it is also not useful, I am not in a buying mood, and if I want to buy something I don't want to have non profit sims. I also looked often for services from builders, scriptors, job agencies, rentals and so on. This category drowns completely in SL. And there should be a report system with consequences if someone abuses such a system.

Another approach could be a juror team that visits the places, shops and so on and gives neutral ranking. That would really help to find quality. But I guess it is very difficult to realize.

The fact is, they hardly have any impact on the stability of the grid (I could go in technical details about it, but I feel this it is not necessary). Another claim they have is that their the traffic system is easily gamed. And it is. But it is every other system you put in place to replace it. Google ranking system is their trade secret, nobody except Google knows how it works, so you can only guess by trial and error how to game it. And it did not take too long for people to figure out. I have seen places with traffic of 0-50 ranked #1 on a single keyword search.

1. The recipe: split up your land into small parcels, down do 300-400 sqm meters.
2. Litter it with objects marked "show in search" and contain nothing but spam words in their name and description (funny thing is you must not have a lot of these, 10-20 at most, big places with hundreds of real objects for get ranked a lot worse).
3. Have just a few profiles have it in their picks
4. Profit by gaming the system by getting rank #1 for totally useless place.

Now I have seen this happening with increased frequency, I am monitoring a few key words I am interested in myself. What I am seeing now is that for most popular searches 3-4 out of top 10 are gamed and totally useless results. This percentage is increasing daily. My guess is that the new search will become totally useless in no time.

The old traffic based metric has worked well for so long, you skip the first few most obvious offenders and you start finding really what you were looking for. None of this black magic totally obscure way that not even LL knows exactly why some results come up in a certain way. The traffic system did create some problems, but they are exaggerated by the "activists", and it did bring a lot of good to SL, became part of the culture, and allowed new users to taste what it is like to have a few lindens in their pocket and enjoy their SL experience.

However, it's still annoying (and was why I created VWR-7617) that the places that you find in Search -> All give you no way to access the map, short of redoing your search in Search -> Places. This is cumbersome.

I suspect that the final solution to this will involve actually fixing the HTML search so that it also gives a map link.

The reversion of this tab (Search>Places) was in response to feedback directed to the Search team, headed by Jeska Linden. Many good points were made as to problems with the "new" layout we introduced in 1.20(7), compared to the "old" layout that has existed in previous versions.

For this reason, the Search team decided to leave Search>Places to be the same in the 1.20 viewer as in the 1.19.1(4) viewer. This does not imply the team will revert this permanently, or all Search tabs.

By contrast, the Search project does intend to make future improvements/progress on Search; but it was simply too early/sudden to do this in viewer 1.20 without more opportunities for feedback.

In her blog post http://blog.secondlife.com/2008/04/28/second-life-showcase-popular-places-and-the-future-of-traffic/ , Jeska Linden does invite interested Residents to join a specific Second Life group to learn about any upcoming brainstorming sessions to be held inworld on these topics.