1. We already have a function on the land menu that enables people to be banned for "no payment on file" or "payment on file" but not used or even "payment information used" – all three categories. I don't use any of these precisely because any of them could bar paying customers from my rentals. Some people associate NPIOF with griefing or alts so they ban them – it's the Lindens' response to the outcry over the 6/6/6/ decision to allow free accounts without information on file.

2. I'm suggesting that bots be identified as bots. That if you want to operate bots in SL, you go to a sign-up page, and, as a premium, paying customer, you pay for an account that is marked as such as a bot. There might even be bulk discounts to enable companies to buy fleets of them. But at least they will be paying for this use of scarce resources, rather than taking it for free, and they will also be marked as such with this type of account. It's very easy to do. Add to the land menu the ability to block them – and you have solved the bot problem amicably for many, many people, while still enabling people to use them if they wish.

3. You're not making sense with your claims about estate management bots somehow being "discriminated against". Why should you be allowed to have accounts for free to perform automatic functions for your land business that the rest of us who are not scripters, do not have the means to pay scripters, and do not have the means to maintain servers and web pages do not have access to? You should pay for the use of these resources as a higher-use account. If anything, the discrimination occurs by the use of bots for free by those with programming skills and resources to maintain third-party websites that they can use to exploit the economy. The system cannot distinguish between land botting and estate management and it doesn't need to. There is nothing wrong with land botting and estate management by bots. But it should be marked and paid for just like I have to pay for my tier-bearing accounts for land.

4. You don't need an "override" as the suggestion is to have optional check-off boxes, just as we have now, for NPIOF. If you don't bother to check off the box, the NPIOF will come freely on to your land. In the same way, if the system defaults to no checks, as it does for NPIOF, you'd have to manually check off to block bots. I think that's fair.

I would also propose that, as an incentive and for general benefit, there could be different levels of avatar presence for bot accounts. If one has a landbot or a bot which takes photos of people on a fairground ride, it doesn't need the same abilities as a real-person avatar, it barely needs to have a physical presence at all - so why bother lagging the sim with any rendering and presence tasks that it will never use? Cutting out the unnecessary functions

There is the practical aspect of how all of this could actually be enforced of course, particularly if there is a charge involved, which would provide a great disincentive for people to register. (Charging is not something that I am in favour of as a blanket measure; it should depend on the resources needed. A bot which logs in once every day for a minute to handle some irritating group function puts less strain on the grid than having a human doing the same would; a bot which hammers the search and the asset server crawling the grid twenty-four hours a day is a different animal.) It is an important point, though, and just letting anyone log in with anything regardless of what it is is not a solution, it is avoiding having to make decisions.

I personally do not allow camping on any of my properties because I think it's a waste of my money, and the newbies' time. I think most people can be encouraged to give up a latte and buy $1000 Lindens for US $3.65. If they can't do that for some reason, they have money trees, contests, small jobs, etc. they can do to raise money.

It's not that charging would be a disincentive; paying would be the only way you could get such an account.

I'm also not emotionally swayed by people trying to appeal to the higher aesthetics by saying bots are used for art and acting. They can pay too. Their use is a minority use. If a person is capable of managing a bot, they are already someone with education and usually resources, often institutional resources. They can pay $9.95/month as owners of 512 land parcels do – or perhaps some other reasonable fee, $3.65/month.

The only people seriously discouraged by any sort of fee and marking are those using hordes of bots for traffic boosting and vaccuming of camping stations. I'm not for continuing to give them a free ride, a free use of resources, and a free boost in a competitive economy.

Land dealers, artists, scientists – they can all afford to pay and should pay.

Again I can't believe it but I am agreeing with you again, they can pay and should pay and the best way to at least make a barrier to entry (other than coding skill) and at the same time compensate LL for the increased resource usage, which will presumably go towards grid upgrades and such to help offset such usage.

But if one is suggesting a charge based on the waste of resources that bots represent, one shouldn't charge a bot which logs in occasionally for a minute the same rate as a bot that is constantly on and teleporting and making search requests. SL doesn't charge a flat fee for builders regardless of how much land they want.

I think that in practice the solution to your point D is broader and better LL regulation on the gathering of personal information, as the technology used isn't that significant; plenty of data can be gathered via LSL already and that needs to be regulated just as much.

Where is this new CEO anyway? Already quit in horror?

go right ahead and wax idiotic about his: It'll end up closed by a Linden or a dev and marked as being impossible to implement at this time.

There is ENORMOUS sentiment in the general public at large AGAINST bots. This feeling isn't held by engineers and programmers, particularly the tiny percentage of those who are active in policing the JIRA. But such persons cannot be allowed to dictate the terms of Second Life, when so many are impacted adversely by bots – their privacy is lost to market scrapers, their money intended for newbies is sucked dry, the performance of their sims are degraded, and overall the grid is strained by hordes of artificial creatures taking up space and energy.

The notion that you cannot distinguish bots from regular log-ins is ridiculous. Of course you can distinguish them. Bots run on commands – these commands have to come from other servers, and can be detected in patterns of activity. There is no organic, random conversation and response to the environment like a human among these still very primitive AI entities. It's another matter whether reading this pattern of activity is too time-consuming or resource-intensive.

Software is used to run these bots – it has a pattern of activity. I'd like to hear a greater variety of sources on the question of what is involved in reading and interpreting such commands and such actions.

Furthermore, the Lindens can simply decide as a policy that they wish to require all persons using avatars as bots must mark them and simply enforce it by urging compliance, and opening up abuse reports from those who are concerned about the patterns of behaviour of accounts evidently being used as bots.

Everybody can tell a bot from a regular person after a few minutes of observation as I've indicated.

Please stop closing JIRA proposals for features you don't like. There is no reason to do so and if you do that, they will be re-opened. Both Universal and Gordon have a pattern of protesting and closing proposals from me that they don't like, and they should cease that sort of harassment.

On the topic of closings, it is you're opinion Prok that only Lindens should close it however the lindens have never confirmed (or to be fair denied) that notion directly however they actively request people help out on the JIRA and I won't even bother linking to it since it's linked to you every time you pull one of these hissy fits, grow up Prok this is the real workd where not everyone agrees with you and you aren't always right, either that or ask a linden to close you're account and wait 15 years until you're old enough for a tg account since that's about the age you act when you get all fussy on the JIRA.

On the topic putting the job of bot patrol on every resident is a huge problem because it would lead to a huge number of false positives as I'm sure you'll even agree the residents here can't really be trusted and also previous experiences from LL doing this with gambling, with ad farms, etc... have led to only limited success because most residents don't really care unless like in the case of ad farms it becomes a large annoyance or they're just vindictive against anyone and everything like you in which case I'm just guessing but I'd guess that LL just ignores their reports after awhile and that includes AR's and JIRAs incidentally.

So carry on with the abusive commentary. Why not see if LL can apply the same even hand of justice hand all around shall we?

Just because you don't like prok or think you have some god like powers because you can close a pjira (anybody can so you are not special) doesn't mean you should. If you don't like something then why not make a logical and mature comment about it and leave it at that.

After all... It is not up to Universal or Gordon or any non LL staff to decide if something is possible or not possible. I recall people saying solid surface light occlusion was not possible in SL. I.e.; shadows. Dark interiors, etc. Yet LL has a programmer working on it and achieving some degree of success. So it seems all those who arrogantly snorted "it is impossible" were just... ignorant and wrong.

Let LL decide. SL belongs to LL and LL decides what they will or will not do. End of story.

As for on topic... I want NPC mobs. There is a use for bots. LL can implement something to allow identification of bots but it is unlikely to be something they are interested in investing in at this time as it would require a registry. So leave it to LL to close please. This has as much merit as the arguments in favor of bots running loose performing experiments. It is just that anarchy is easier than governance so what do you work on first? LL's call. Not yours.

The notion that you "can't tell a bot from a human" is, again, false on the face of it. A human being observing a bot for a few minutes quickly determines that a bot is not human.

– it always lands in a hunched position
– it doesn't answer IMs
– it doesn't buy anything
– it doesn't build
– it doesn't talk
– it doesn't walk around purposely, stopping randomly to look at things, but sort of floats just above the ground

The Lindens monitor every single transaction of ours and every single word we type. They may not look at these things but their servers record it for their perusal. So they have the option to look at this information and determine what is a bot should they desire. Precisely because the operators of bots have to run them by remote control with commands, they can simply scan for those series of commands.

"Trolling" is not a concept I recognize. It's far too often abused and misused against people merely holding a minority opinion which is not liked by some group. In this case, sentiment against bots runs very high in SL and if put to a vote, would likely wind up with the overwhelming majority of those who voted being even for banning bots, not simply regulating them as I suggest. I fail to see why a determined minority of coders on the JIRA should keep determining the features of our world.

I don't endorse the Lindens' system of allowing "helpers" to determine what needs closing as I see how these "helpers" behave – hysterically, arbitrarily, meanly. I think this thread is a good illustration of who needs to grow up.

Should a Linden come by to close a feature request ... well then, the administrators have spoken. Deal with it.

IMO though, the bigger issue for LL would be the arms race this would start and they have already said that they do not want to get into an arms race. For instance, say they look for the series of commands going to the bots, the bots would simply be changed so that their commands are encrypted and based on the date and time so that the command would look different every time. Now LL has to combat that and so on.

My proposal is legitimate, and frankly, even have no call to close it. It's a feature request. Many feature requests seem "impossible" until the political will is gathered – how about the policy developed on ad farms, to cite but one very crystal-clear example of an action that the Lindens finally took as a policy, not an engineering feat that came after a lot of community discussion and proposals.

I don't think "arms race" is a valid argument. The Lindens face arms races from those trying to create casinos under the radar – there are no shortage of them. They face arms faces from those griefing – no shortage there. They face arms races with "ageplayers" hiding child pornography – and with all of these challenges, they have dealt with them with a mixture of policy and tools and abuse reporting. They don't duck a challenge just because it has no neat technical solution to fix it.

Lindens rarely chose the easier path for themselves when it comes to certain undesirable phenomenon. For example, they could simply deprecate – not allow to function – a few of the major casino scripts and that would be the instant end of the casino business, without a single abuse report having to be fired. They could deprecate various other weapons scripts or mega prims used for griefing – they don't. They never work it that way.

You just can't deprecate a script, they would have to deprecate LSL functions that the scripts use, doing that would break thousands of other scripts that have nothing to do with casinos or weapons. And even if they could somehow disable casino scripts without effecting others, then it would start an arms race.

The Lindens have shown collossal indifference about the breakage of resident scripts. Each patch that comes in breaks scripts all over the place. I now routinely have to reset certain rental boxes that freeze each time the Lindens do a new patch or restart a server – and certain things like those pets that follow you just get broken for good. The blogosphere is littered with inventors wailing about their broken scripts. When the Lindens want to break a script, they break a script. Our God is an Old Testament God here. There is no Redemption.

Universal, you are not a valid interlocutor.

Harleen, the Lindens did stop Copybot – by policy, not technology. They made it a crime to use it under the TOS. The same can be done with bots in general unless they are labelled.

Wake up and face both facts and yourself Dearie: Ignoring the truth does not make it go away.

If bots are supposed to use some log-in parameter – let them!

Don't dictate to me what I can and cannot say or what does or does not need defending. Cumulatively speaking, bots use up more resources. We're seeing this in spades now with traffic boosting and with the draining of camping. The Lindens have now monkeyed with the traffic and essentially removed it from the sorting of search/places – something I rallied against for months in advance, with many people saying "oh, they'll never do that!". But they did. And it won't make any difference to the use of bots because while traffic is removed from sorting results, it's not removed from the parcel or from search completely as a factor.

I fail to see why bots which gain advantage for their users in the market in the following ways shouldn't be regulated, labelled, and charged for:

o theft from camping sites
o land buys and sales
o traffic boosting
o market data scraping
o scientific experimentation for experiments usually with substantial institutional support

Why should any individual gaining any of those commercial boosts get a free account, while others who tier land, and have to suffer the reduction of their land value and use have to pay and thus sustain these exploiters?!

THAT is why they need to charge. Otherwise, WE are paying. That is injust, and not a free market economy, but socialism, on behalf of a New Class.

And I did not also mean to imply that just because things work for me they do not for someone else, I also have never had a problem with my rental scripts that doesn't mean I think you have not. And whether they stop functions from working to prevent things like gambling does not matter people will adapt around it. Say for instance they stop payment to object functions from working, this would stop all vendors from working also, store owners would adapt and go back to selling from boxes, renters would adapt and sell rental boxes instead, underground casinos would adapt and hire pit bosses or create NPC ones (oh, the irony!). Now, LL is in the arms race they were trying to avoid.

Surely the commands that operate bots can be broken down into various functions. There may be certain commands or strings or whatever that are unique to bots, and these could be blocked, for example. Once you open your mind to entertaining all the possibilities, they will avail themselves; it's the ideological hammerlock that prevents solutions from coming.

Once again: You are barking up the wrong tree.

I'd like to hear from a wider variety of sources about the question of how AI systems are automated, and whether indeed a) there is absolutely nothing in their automation distinguishable from a human-operated account b) there is absolutely nothing that can be done about this through spot-checking of known live-human patterns; c) there is really nothing that can be done to enforce in fact a uniform system of bot log-on that would reveal it to be a bot.

All of these areas can and should be explored.

"Bots. Based on a set of behavioral characteristics that we observe bots having, we believe that about 10% to 15% of our user hours are attributable to bots. This has been consistent for some time."

http://blog.secondlife.com/2008/07/08/second-life-virtual-world-expands-35-in-q2/#comment-614159

Furthermore, after establishing that bots are not the "30 percent" that many believe, Zee goes on to talk about increasing user activity and log-ons not being involved in the Lindens' economic metrics; and quite frankly, that raises the question of whether bots make up as much camping and traffic-infusion dollars as is widely believed

"The corresponding metric that we think indicates consistent non-bot activity in Second Life are the user to user inworld transactions per hour and the LindeX volume per user hour. Both of these have been consistent certainly since the bad of gambling at about $0.87 and $0.27 to $0.29 per user hour respectively. Our revenue has also been consistent at $0.20 per user hour. If bots were an increasing percentage of our traffic, I think there would be deterioration of these economic metrics."

Therefore it does no good to allow banning since they are there because the parcel owner wants them there.

Land bots have transactions and move around a lot and thus look like a human controlled account with the exception of timing. And I so seriously doubt anyone at LL has the skills to conduct this type of data mining. Not to mention I doubt some of the key data is even logged or recorded anywhere for use in data mining.

Linden Lab promotes the use of traffic falsification bots by virtue of the fact they do not permaban parcel owners found to be operating them.

The client code would have to be closed source with specialized security tokens for bot regulation to be remotely possible. Mitch Kapor calls the shots so you are barking up a tree that is not going to budge Prok. Linden Lab serves Mitch Kapor. That much became pretty clear the other day. It was his money. It is his company. It is his game. He calls the shots. Accept it. SL is not a democracy and will never have governance intended to serve the customers in some sense of fairness. It is a laboratory running Mitch's experiments. It is an utter waste of time to lobby for "the right thing".

Nice try though. Noble efforts are worthy of being appreciated even when they have no possibility of being realized.

As for published metrics... without access to the raw data there is no means of independently confirming the metrics as valid. Take that any way you want.

All these functions should be available without bots. The major reason is that bots use available avatar slots, which should indeed be used by humans — since they're so limited.

Although I understand the need for some bots, like so many other things, bots get far too easily abused. I'm not for banning them from the grid — since they have legitimate uses — but certainly an option for landowners to keep them away is more than adequate.

As for identifying what is a bot and what isn't... well, "observation" is naturally not enforceable technically. However, all SL clients — and a libSL client activating a bot is also a SL client — will identify themselves as such. What LL would need to do is to allow real SL clients (eg. onRez, Nicholaz, AjaxLife, and so many others...) to register their special identification string on a database (not unlike what Apple does with iPhone applications — for exactly the same reason, keeping abuse as much under control as possible — and Apple charges US$99/year for that!). So basically the checkbox would validate if an avatar comes from a registered, valid SL client (with a human behind it), or just from a libSL-created bot.

Granted, the current system is way too easy to fool (it's very easy to "pass" as a legitimate SL client), so LL would have to design another level of validation, e.g. using an unique token/ID for registered SL clients that is not possible to duplicate. That way you could indeed flag your parcel for use only by "registered SL clients".

The Lindens have stated unequivocally that they can detect bot log-ons. Great. Detect – and label, and charge, and regulate. Good or bad, they use resources, the overwhelming tendency is to use them to traffic infuse, and that should be capped through charging.

I also think everyone should have the right to keep them off their land by banning them generically – they also are used to go around scraping data, giving some people an unfair advantage of they have the resources to scrape and maintain third-party sites, as they grab everybody else's data and use it themselves commercially.

While a token could be devised, you can also just as for voluntary compliance. A special bot account that you sign up for, it need not even cost $9.95, it might even consciously provide 100 bots in an army that you get at a deep discount.

And know this: as long as a client can access SL, people with real bad intent can always clone the client and make bots available. Trying to discriminate bots will only harm innovation and good intent.

I fail to see why I have to pay for accounts that hold mainland and groups, due to the 25 group limit, but you don't have to pay for your "innovative sports". I'll bet your "innovative sports" team goes off course and lands in public infohubs and takes up spaces now and then as many bots do.

There's always someone to come along and say, "But my bot is for Science". There's always someone to endlessly, without any restraint to cry "but innovation and good intent trumps all". This is the endless SL cry. And yet it diminishes the experience for everyone else – what about their innovation?

If you abandon early the responsibility of humans to label – and restrain bots – you are later then responsible down the line for the murder of humans.

The idea that there could be a "witch hunt" is patently absurd. Mechanical functions are easily identified. LL can ask first of all for self compliance by making special labelled accounts and charging more for them, as they remain online longer and use more resources.

Your comments about FIC and lynch mobs make no sense. One the Lindens have the political will to differentiate human beings from bots – and good Lord, they should do THAT much if they are moral and responsibile – then they will develop reasonable ways to follow through, ranging from automatic detection of routines to self-compliance with labelling to ARs.

The constant invocation of witch hunts here is a really troublesome distraction from the tyranny of the coder to arbitrarily create a machine and enforce his will on the rest of the unwitting public. Why is that ok?!

A note about the load from bots. I haven't seen anything from LL stating that bots cause a problem with resources. Bots do not continually ping the asset server for inventory. They don't even get continual updates from the simulator on objects running for rendering unless the running bot specifically asks for it. The client viewer running with a person behind it is a far bigger load then a bot because at times a bot only needs to maintain the minimal connected session if it is idle.

I completely understand the desire for people to want to prevent bots from appearing in their private areas, but the simulator has no real way of knowing ahead of time. Just like a web page does not know if it is being spidered or if a real person is viewing the page. The only real way is to come up with a policy for landowners and such that is similar to the robot.txt, but even then it only applies if the bot owner is responsible and looks for that check.

The only way out of that is to ban everyone from coming to your land unless you specifically allow them. That is how Facebook does it. Once you install a widget or befriend someone who can run a scraping program, all your information can be pulled.

Personally, I do not think that bots should require paid accounts unless someone has a large number of them logged in all the time. Something like that would require special circumstances anyways because the number of alt accounts is limited per person. If someone is violating that and creating lots of bots, you can't expect they will be responsible to create and pay for a special bot account.

Bots can – and are – identifiable. By their patterns of activity inworld. The log-on protocol somewhere has to differ than a human simply because a program is first accessed. That may be out of the recognizable loop, but it is the difference.

Bots are a drag on resources because they are not people who socialize, buy and sell, create, contribute, etc. They just infuse traffic by and large, with the exception of a tiny percentage used for other functions, sometimes group management or RP stuff on theme sims.

Of course bots can and should be labelled. If it is true that OMGODZORZ you can't tell a bot apart from a human, then that's a human choice to make it so. It's not just "emergent behaviour that we can do absolutely nothing about".

This is about human will and intent, and about the human will and intent of some humans trying to have control over others with automatic routines that they themselves code and control.

Asimov was wrong, in fact. The morality of a robot isn't contained in its own coding so as not to harm human beings. The morality of a robot starts in the morality of a coder, and they start with identifying this manifestation of their will for what it is – automatic mechanistic manifestation of their will. The idea that coders will make robots with free will and then hold up their hands helpless that they simply can't do anything, these AI creatures "have a mind of their own" and "you can't even tell them apart" is morally, ethically, and legally wrong. Take responsibility for your code. Stop trying control other people with it just for your own self interests against the public interest.

Facebook does not at all work in such a way as to prevent scraping from everyone by being closed by everyone – it not that all or nothing a routine. You can select to be public or not public. And your information is visible only if you friend another person, otherwise, the information they see is visible. People should have exactly the same opt-ins functions on bots. Bots serve no demontrable public interest, serve only the private wills of private interests, and the public is not required to cater to them.

The first step is identification of bots as bots as a policy. Everything else flows from this.

2. the existance of the registration api clearly indicates that Linden Lab doesn't want the paid members only model. You are living in the past on a platform that is supposed to be about 'the future'.

3. see my response to #1

A. an AFK user might not respond. Also, someone that doesn't like you. People aren't required to respond to you. As for clones of lotharios, maybe there's not a lot of originality. Either way, #1.

B. "Those who chose to use camping or money trees or any kind of legal game to distribute free money should not be exploited by programmers using bots." oil should not be exploited by americans for driving, but it continues to happen. If cows were smarter they would run away, but they seem to like being turned into Whoppers and Big Macs. You cannot mandate human behavior to make people smarter.

C. "I don't see why these free accounts should be allowed to draw down resources" resources used are typically limited because data gathering bots are (supposed to be) designed to be efficient. As for you 'not seeing why', it's been explained enough. I think you're just slow or want attention.

D. Second Life is a SCIENCE EXPERIMENT. Reference: Linden LAB.

Socializing, buying selling, creating, contributing... These are all things that actually create a load on the resources. A bot logged in that is not accessing inventory, and not requesting object data updates for nearby objects can consume a tiny percentage of the resources that a person logged in would. traffic is not a resource, it is a measurement of the number of avatars that have logged in to a simulator. Bots will count towards that, so if a person has a problem with no one being able to access their sim because it is full of bots, they may need to change their accessibility options for the island (explained later).

The best solution is to set a precedent like was done with webpages and robots.txt. But even then it will always be voluntary on the side of the person running the bot code. It is not enforceable.

Stereotyping all coders is wrong, everyone is different, and most coders make wonderful great nice things that everyone is happy with. Making coders out to be some kind of oppressor is silly.

Facebook does work like I said, and there is a direct correlation to SL. If you friend a person they can see everything you have on your profile. Your profile may be not public and that would be the same as not allowing public access to your island. As soon as you allow a person access to your facebook profile, that person can scrape all your data, either with a bot or with a program. If you don't want them to do that, you don't allow them to be a friend.
For SL you would not allow public access to your island, and you would only allow access to people that you know and trust to not be a bot, and to not run a bot using their login info. That would be your equivilent to your Facebook example.

I think it would be better to handle it the way websites do. Add something to the land description that is something like "NO ROBOTS HERE" and the encourage bot creators check and respect that as well as identify in their profiles that the account is used for bots. This could then be used for self identification
I do not see a reason why bots should be required to be premium, and given the suggestion described, it would still not be enforceable.

I think this, identifying Bots as such, and adding flags to ban, sale, some script functions and where ever appropriate would solve most of the issues raised. And if the Lindens state they can detect Bot logons, set the Bot flag automatically. Bots ain't people and they should have no right to object.

However, I do agree that making you pay for a Bot account would be a bit harsh.

if the Lindens state they can detect Bot logons

From my technical point of view, this isn't possible (for all cases). Even if LL came up with an algorithm that could detect bots today, tomorrow would be different and it would turn into a pointless cat and mouse that would take away from other uses of time. There is no way to guarantee detection of a bot before it performs an action (such as land grabbing). I think it would be better to solve issues regarding land snatching bots through a different method (timeouts and/or sale to specific individuals).

Allowing no bots on your land is kinda paranoid, i don't even see what a bot would go do at your place unless its to test out the performances or index your parcel for good reason.

Ok bots that create traffic are not so good cause they cheat the traffic system in place, but LL has made policy changes regarding this and they are taking action against the ones who abuse the traffic system.

So please chill out, i really don't think there is a major problem with bots, there are here to help the Lindens most of the time and they are nothing else then a program created by skilled peoples.

Going to their level and making good logical cases for why something is possible and something not is in those situations pointless. Sometimes it's more efficient just to ignore those proposals rather than arguing there or resolving as "won't finish", cause gathering attention is the actual goal of the starters of those proposals sometimes rather than clarifying issues.