Bye bye traffic bots

When I got that mail from Microsoft, I was about to report it as spam, but instead I did some checking, and they ended up flying me to Seattle and putting me up for a week so I could attend the conference, on their dime. So I guess maybe it's a good thing that I don't use your definition of spam.

Well whatever the scenario...LL are awfully slow in enforcing their policies. I still see a ton of bots....also seen places that had removed them, placing them back again.

LL policies are a joke!

I'm sorry, but that's just plain nonsense. However, *you* are free to think of normal avatars that are afk as bots if you wish.

According to your idea, my traffic bots were never bots except when I left my desk, and your landbots aren't bots either unless you leave your desk, because in both cases we have keyboard control of them with openmv clients. We are both able to control them individually with the keyboard that is right next to us. In other words they are manned so they are not bots. That's nonsense, of course.

I haven't differentiated between avatars that are logged in on viewers and those that are logged in on something else - you seem to think I have. I have 6 demo models in the store - bots. For most of their existance, they were each logged in with the standard viewer, and for most of each day their keyboard (1 computer) was right here on my desk next to me, so they were not even afk, but they were still bots. It's not the client that makes an avatar a bot.

I disagree with that completely. Imo, the reason Jack said that they will monitor the search results, rather than have people AR bots, is so that they can control any effect that the bots ban might have on the concurrency numbers. In other words, they intend to go at their own snail's pace and keep the concurrency high - as we've seen already.

There is no need whatsoever to monitor the search results. They create a numerically sorted list of the top 50,000 traffic places every single day. They only need to start at the top and work down, checking whether or not the parcels are set to show in search, and they have all the information they need. All they need to do then is find out why the places have such high traffic - just as they would have to do after finding them through monitoring the search results. The idea of monitoring the search results is just to divert people's thinking.

I had an invitation like that some years ago - when they were setting up their own search engine. They invited a few well-known seo people to Seattle at their expense. I would have liked to have gone, but I don't fly unless it's absolutely essential, so I didn't go. In fact, I got it twice - I deleted the first one with the rest of the spam - I didn't even notice it. The second was to check if I got the first one.

If a manned avatar is a bot and an unmanned avatar is not a bot then what exactly is your definition of a bot?

Also, whoever it was trying to argue their point by comparing bots with spam please immediately stop. Argument by analogy is always, /always/ flawed.

Just looking at Traffic would certainly catch the mass-botters over all terms.
Over a certain traffic count, you just *know* that it's being gamed.

However, I don't believe that LL are even using Search.
Just that example of "Furniture" in Places is throwing up the same 80+ traffic numbers for the same places.

Are the owners LL employee alts or friends?
Have the owners slapped a restraining order on LL?

For the sex places listed under that term, one might consider that the entire future live avatar population of Ursula is currently crammed into a smallish number of high-altitude builds.

The lag in there must be something awesome.
Despite the lag, it must be mid-blowing sex.
It's so good that they have to use horizontal sim-sized megaprims and security orbs to keep the numbers down in there.

I've just had my ass banned for a few grillion minutes in one place
Dammit! I soooo wanted to have it sloooooooow with a white cloud in the middle of a seething mass of grey.
Maybe I should make a "slow bonk with a white cloud in grey' and sell it.
It seems to be very popular.

I don't have a definition of a bot. I'm only saying that yours is wrong. You base your definition on whether or not there's someone at the keyboard, but your own landbots show that to be wrong, because you have their keyboard right there with you. You can control each of them with the keyboard that is next to you, so they are not afk, and yet they are still bots.

The thing that makes an avatar a bot is that it is programmed/designed to perform certain functions, even if the only function is to stand in one spot and say or do nothing else (e.g. traffic bots and store models) and it will perform the function(s) without the need for anyone to be operating the keyboard. They are usually created specifically to *be* bots. The client they use is irrelevant.

Normal avatars are not programmed or designed to perform functions. They can become botlike (a word you used earlier) but they don't become bots. For instance, in a little while, I'll go into another room to watch the F1 Grand Prix. if I (my av) were logged in at the time, I may not log out, but I'm not programmed or designed to perform a function, so I won't become a bot while I'm watching the race. I may be botlike for that period of time, so that you are unable to tell the difference, but I won't be a bot.

If you really want a definition for bots, perhaps you should be thinking along the lines of "avatars that are programmed or designed to perform automated functions".

I agree with you. Imo, it was just something to say that would allow LL to go at their own snail's pace and yet satisfy the majority of people.

No, in THAT SPECIFIC example, what makes it traffic gaming is the person SAYING that it is traffic gaming. Without the admission, it is an edge case.



Doubtful, because there are too many legitimate uses for "one or two bots", and the traffic bump one gets for one or two bots isn't nearly as significant as the more blatant abuses that they are going to focus on. They aren't likely going to go down the list several pages to find mom'n'pops store with a model bot that secretly is bumping traffic.

You also don't know that they won't check what someone says in the forums. It's not like it is very hard to search. Evidence is evidence; *I* would.



Firstly, I was using a euphemism, much like "if you don't pay your taxes, the IRS comes calling".

Secondly, with this policy, LL IS "calling" people and not banning them on sight. They did the same thing with the adfarmers and extortionists when they announced those policies, too. They informed the most obvious and blatant abusers, and gave them a certain amount of time to clean up their act before they resorted to taking their land or suspending them. Suspensions didn't start day one after the policy announcement.



So? Where have I disagreed? Why do I have to keep repeating that you're preaching to the choir every goddamned time? I'm well aware that they've screwed up on SOME edge cases, and that they have a history of either 1) complete inaction, letting things go to hell, or 2) screwing up with too vague or to specific of a policy and muff enforcement because they don't train their staff in proper enforcement techniques. Old news.

Even still, what we have now, with respect to the adfarms and extortion microparcels is a DAMN SIGHT BETTER than it was over a year ago. Is it perfect? No, but after previous attempts at policy failed, they finally got something close to being right, and we have the positive effects to show for the effort we put in.



I didn't; that was your ASSumption.



OK. I went and looked at said "adfarm" in the NW corner of Dangsan. There is ONE billboard near the sim border, a tiny Arbor project plaque a little farther down, and a cube ad a bit farther down than that. Quite a ways farther down, almost middle left of the region, there's another of the same company's billboards as the one close to Coonspiracy's border. THAT actually is a violation which can be ARed and rectified.

Across the road from the adfarm, someone has a business there with a big 10m cube floating 40m above their store. I saw your giant invisiprim screen there, but there's nothing even remotely that big that the screen is covering up.

That's all I could find. So, what, exactly, is so "worthless" about the policy with respect to what little is there? 3 billboards, only one of which is anywhere near your property, and it conforms to the rules for size and ground attachment, plus an illegal redundant one which you could easily AR and have removed.



Neither do I; that's why I'm glad they are changing their tact to one which they actually have to balance THINKING with POLICY for enforcement. In the past, they get too caught up with formulae and detail, and end up either not acting at all (which fucks up innocent people from inaction), or they act on the wrong problems which were often not problems at all. Why do you think Carl got ganked? It certainly wasn't because the person who handled his case was THINKING about the problem; he/she just went "oh! Nipple in PG! *BAN*".

So, yeah, I WANT them bothering to research things like intent, ESPECIALLY on edge cases. I WANT them to *THINK* before they smite with the banhammer, and I WANT them to implement good policies and train good folks to handle themselves well.

I didn't get everything *I* wanted in the adfarm policy, either, but I am happy with what I did get from it, enough to stomach the compromise.



Why are you being so antagonistic if you're not disagreeing with it?

Is that really called for?

I'm happy that it worked out for you.

NOTHING in that would make me believe that ANYTHING Microsoft sends me UNSOLICITED through email is anything other than spam. They could offer to PAY me to visit, and I would STILL throw it in the trash.

I don't have time to waste on their crap.

Well, you can believe whatever you want, of course.

They did the same thing with the last adfarm and extortion policies, too. We busted Jack's chops weekly whilst they gave the parasites weeks to conform to the new policies. However, eventually, they followed through. Ads disappeared. Extortion microparcels were confiscated and swapped.



uhh.. if they pull a regular report which IS a numerically-sorted list of the top traffic places every single day, that IS "monitoring".

I'm not sure why you are disagreeing with that, because you've basically just described the same thing.

If it plays out the same way that the recent policies have, it will be another week or so, and suspensions will start showing up on the incident report blotter. People will go back to not logging in their bots because they'll see LL is serious.

As can you, of course

No they didn't do the same thing with the last adfarm and extortion policies. They didn't say that they didn't want ARs. It's totally different.

Actually no. Firstly, it isn't monitoring anything. It's not something that they look at. It's *not* a report. The list is created for something that has nothing to do with monitoring. Secondly, it isn't monitoring the "search results", which is what Jack said they would be doing - the list doesn't come from the search results.

That may well be, but it's something that I'll believe when I see it.

bonus points for the Frantics reference.

"Approach, students. Close the circle at the feet of the Master"
TQL FTW

OK, make it Apple, or Red Hat, or Linden Labs, or whatever company you might like a free trip to their HQ for a conference from. If they send that invite to 10,000 people, that's spam. If they send it just to you, it isn't. Doesn't matter what the content is or who the sender is.

Unsolicited commercial email still isn't spam UNLESS IT IS BULK, and a for sale sign on a parcel is not adcutting UNLESS IT IS A MICROPARCEL, and one or two bots in a sim aren't the problem that Linden Labs needs to be worrying about here. It's better that they miss some marginal cases, even if they obviously have fraudulent intent.

I'm not saying that the specific case does not have fraudulent intent, I'm saying it's not something that Linden Labs can tell just looking at the facts they can see in world.
And there's still a multiply-ARed adfarm sitting there, right next to our group land, because they defined the rules to allow them to continue. Yes, those ads are legal under the new policy. THAT ITS THE POINT. Their definition of an adfarm doesn't include microparcel ads "that are attached to the ground and close to the ground", and it DOES include ads that aren't on microparcels. It's ATTACKING THE WRONG PROBLEM, the ads, when the real problem is the microparcels.

If you think they should be going after edge cases that they can't tell from legitimate uses, it's just going to hurt innocent people and divert resources from real problems. And, christ, do you honestly expect the G-Team to search the forums to see if someone was boasting about pushing the envelope? You say that's just normal investigation. From the G-Team? REALLY?
But you know how they really interpret policies, so the best we can hope for is they define policies that don't start their investigators down crazy paths.

I tend to get antagonistic when people seem to be deliberately misrepresenting what I'm saying.

We'll find out if LL is willing to knock a very large percentage of the concurrency numbers offline. My prediction? They'll go slowly enough to give bot makers time to make "better" bots, so their grid doesn't look so ridiculous.

When bots are walking, flying and talking gibberish, like they do at clubs, LL will call the policy a success, but all they will have accomplished is to make it impossible to count the bots.

That's the plan.

It never fails to amaze me that LL routinely turns policy enforcement into the halting problem.

By their glacial standards it's far too soon to tell, but one can't help wondering if LL really ever intended to enforce this one at all. I mean, if they did, it still wouldn't have any appreciable effect on Traffic's utility in Search, but rather merely change the "game" to a competition for how closely one can approach whatever turns out to be the enforcement threshold using whatever turns out to be not-quite-bots.

They've already gotten all the benefit they could have wanted from the policy: it's given the anti-botting brigade a win. Their moral high-ground has been confirmed with Lindens on the record that some practices are too distasteful even for Second Life.

What could they possibly have to gain by actually trying to enforce it? It's not as if anybody's bonus is contingent on decreasing concurrency.

This, of course, is the second time that they've made changes to the Places tab. On the first occasion they replaced traffic rankings with All>Places rankings, in keeping with their own roadmap for search, but then they quickly made an excuse and changed it back again. This time they've announced a ban on bots that intentionally increase the traffic figures, and the way they've done it, it's looking like it will make very little difference to anything. A pattern?

Ohhhh YES THEY DID! At least with the extortion policy (the latest one dealing with Microparcels), Jack told us to hold off on the ARs, that they had a report of the biggest offenders and would be talking to them directly first to get them to pare down their holdings. We asked Jack every week if it was time to start ARing, and he kept saying "just give us a little more time". Well, some of us did; others started ARing when we felt "it was time". The result? Microparcel barons disappearing, dropping their land holdings, prices, etc, and LL swapping out the recalcitrant ones.



It isn't considered "monitoring" to repeatedly pull the same information over time? Interesting. What do you consider "monitoring", then? It isn't a report? Really? What would you consider a "report", then? It doesn't come from search? Where does it come from, then?

You may be using your own definitions of words, and that's fine, but I think you're just setting up for a silly semantics argument, myself.



When it comes to LL, skepticism on the part of its residents is a healthy thing, since we've been given ample experience to bolster that feeling. However, at least to some positive degree, they've actually been accomplishing things they have set out to do, however imperfect the results may be, on the whole.

Yeah, your definition, and that's fine. I'd still throw it away, UNLESS I SPECIFICALLY OPTED IN FOR IT. I don't have ANY desire for "free visits" to corporate HQs, unless I have specifically indicated such to them already, and that list is haltingly short.



Did you mean "adfarming" (not "adcutting"?. I am ambivalent on the issue of whether a sim cut into 512s has a justified need to have a large "for sale" sign on EACH and EVERY parcel. Personally, I think it is rather redundant, but I have already admitted that it was an unnecessary side case that the adfarming policy didn't need to address. I am willing to take that little bit of collateral damage (so were the various land barons who were also fighting to get rid of adfarms; just to avoid that argument), so that the greater good could be achieved.

No, you're right, one or two bots in a region shouldn't generally be a problem, and the stated action in the policy won't affect them for some time, if at all. However, if someone sticks one or two bots in a box high over his land, and it gets ARed, then LL should respond to the AR and investigate. If they find that the bots do serve to inflate traffic and serve no other useful purpose (chat logs are great for that, doncha know, as some people just don't know when to keep their mouths shut), then the owner should be sanctioned for it appropriately. I certainly don't want them to ignore blatant and excessive violations in favor of small-time abusers, but when the majority that is left are small-time abusers, they still should deal with them.



I dunno. Two bots dancing in a club would tend to be questionable, to me anyway. What other point is there to their presence? Sure, it can be argued that they could be group inviters/greeters, but TWO are needed for those functions? I think that the facts they can see in-world would generate justifiable suspicion; at least enough to warrant further investigation.



No, but it DOES include "more than 50 total and no more than one per sim", which that billboard DOES violate (and, yes, we've counted the billboards put out by that particular advertiser, and he's been reported for >50 as well). If you want, I would be happy to organize a "flash mob" AR party to get rid of that extra billboard for you. It is silly to have to do, but it does seem to get LL's attention over ones that have fallen into the cracks.

As far as attacking the wrong problem, you're partially right. Ads were most definitely part of the problem, both in number and presentation, across the grid. When we started campaigning, we tried to get LL to enact a complete policy solution for adfarms and their associated microparcels. Apparently, it was too big of a job for Jack, so he opted to do it piecemeal (which we also repeatedly warned would fail to solve the core problem). Yeah, it didn't get rid of all the ads (which I campaigned for with my policy suggestions), nor did it get rid of the microparcel extortion and cutting, so he had to add another policy to complete the job (the recent microparcel cutting policy). Still, I think the results of both policies speak for themselves. You won't find crap like this much anywhere anymore.

I don't know what that adfarm south of you used to look like, but right now, it looks pretty damn nice, save for TWO ads, one of which is in violation, and the other which is barely noticeable from the road, let alone from your property, through all the trees planted by the owner of the majority of that land now. That land looks like it used to be a checkerboard nightmare, but the vast majority of it is now owned by one person who has a green thumb.



Did I say that? You seem to get upset when your words are misconstrued, so why are you doing it to me?

No, I never said that they "should be going after edge cases", but if "edge cases" are reported or discovered to be problematic, then YES, they SHOULD investigate. YES, they SHOULD use all investigative means and methods at their disposal.

When I was in that same governance role in another MMO, we most certainly used forums and in-world chat logs as evidence, and would do searches to gather evidence to help us decide a matter. To NOT even TRY to do so, and then try to render a fair decision was, to us, just dumb. I get the point about the current instance of the LL G-Team being incompetent in that regard, but I am not talking about the instance, I am talking about the principle. I EXPECT them to try and do a good job. Whether they do or not, or consistently do or not is immaterial to that expectation. I ALSO "expect", just like you, based on experience, that they won't, but that doesn't change my base expectation that they SHOULD be TRYING to do a good job. If I didn't have that as a base expectation, I wouldn't be here and in-world at Jack's giving them shit over it. I simply wouldn't care.



I know how they have interpreted policies, and it may be fuel for my low expectations of how they will interpret future policies, but they are not robots. They are not static. They CAN change. Whether they will or not is subject to debate, but what is not subject to debate is that they CAN. Hence, the best *I* can hope for is that they do two things: 1) Make better policies, and 2) Enforce them better.



Well, then stop doing it to other people, k?

That's your choice. It's still not spam. It's something you don't want, but if it's not bulk it's not spam.

I'm not ambivalent at all. They don't need to. But that's not the edge case. And that's NOT the problem referred to as "adfarming" or "landcutting" or "microparcel extortion".

There you go again, bringing in something that's not part of the edge case I'm talking about.

Decoration. That's all they need.

There's more than one billboard in the adfarm.

Yes, and they had "impeach bush" on them. The purpose wasn't advertising. The purpose was extortion. Advertising was an accelerator, but not the behavior itself.

Yes, I've been working with the guy who owns that land. I put him in touch with the arbor project. He shouldn't have had to spend all the time and money he's spent on it to reduce (and not eliminate) the problem. That he has is why I say the policy is attacking the wrong problem.


But you know they don't.
I'm talking about the G-Team that exists. The G-Team that's been pretty much unchanged since they were whatever they were four years ago when I started in on Second Life. You can't separate policy from implementation until they change the implementation... and there's been no signs of that.

If I misunderstood you to not actually be arguing that Linden Labs should count edge cases as potential abuse, I'm sorry, but that still seems to be what you're saying to me.

Jack said to hold off on the ARs until they'd had a word with the most significant adfarmers, which meant they would take ARs soon. In this case, they have not indicated that ARs will ever be desired about bots. They've indicated the opposite. Jack said that they will deal with it a different way - by monitoring the search results. So the operation of the bots policy really is different to the operation of the ads policy.

No it's not considered monitoring. Whether or not it's monitoring depends on the use that the list is put to and, in this case, it isn't ever for people to see or for a programme to evaluate or anything similar. Hardly monitoring then.

No it's not a report. It could be a report if it were meant for people to see but, since it isn't meant for people to see, or even for a programme to evaluate or anything similar, it isn't a report.

You seem to be making out that you know an awful lot about it, so surely you know where it comes from. But whether you do or you don't, it does *not* come from the search results, or from anything to do with search.

No I'm not using my own definitions of words. I just happen to know what the list is, whereas you don't. I am using knowledge, whereas you are arguing from your imagination - from your lack of knowledge.

It IS still spam to me, and gets treated as such.



It is an edge case in the overall advertising policy, because some land flippers liked to take it on themselves to put farms of 10-20 spinning yellow "for sale" towers, or huge, spinning, glowing, particle spewing monstrosities of signs on their 512sqm parcels to make sure people couldn't miss the fact that they were "for sale". No, it wasn't the problem referred to as "adfarming", but it WAS a problem with "advertising", and likely why the policy ended up covering them AS WELL AS the ones on microparcels. I would even go so far as to say that said land flippers were taking advantage of the "nuisance" effect to try and get their land sold. That they didn't make an exception for 512sqm+ parcels with SOME of the restrictions is unfortunate, but the collateral damage from it wasn't significant enough to get worked up about.

Again, no, I don't miss them one bit.



If you were simply talking about a single specific edge case in a vacuum, then you'd have a point, but you're not, and I haven't been, thus, I am no more "bringing it in" than you were to "bring it up" in the first place.



I don't quite see the justification there. If "decoration" is all they need, then it becomes easy to justify 2, 5, 10, 20 bots for the same reason. They are there to make the place seem "lived in", for example.



There is ONE billboard, and ONE adcube. The other billboard (the dupe of the one adjacent to you) is over 75m away further down the road. The only other thing that you could REMOTELY call a "billboard" is the single tiny 1m Arbor Project placard which advertises THAT 16sqm PLOT being available to your neighbor for free, if he wants it.



..and there were "advertisers" who went overboard and put out too many ads, or made their ads too much of a nuisance. There were several who didn't have their plots for sale throughout the entire debacle, but they were no less a part of the problem.



He hasn't HAD to do anything. I didn't pay the extortion prices for the microparcels in my region. I negotiated them down to reasonable amounts, or waited them out. Now, I have only two billboards left in the region (and only recently put up, too), but they are hardly causing me any problems, and LL has done well enough to get rid of the giant green particle-spewing towers of doom where they used to be.



No, actually, I don't, and neither do you, unless you're secretly on the G-team. Now, we both can INFER that they don't from their behavior, and it is likely a valid inference, but there is no part of it that amounts to knowing.



I'd say there have been signs of that. SIGNS of that, not IMMACULATE HOLISTIC REPRESENTATIONs of that. The adfarming and landcutting policies have demonstrated some pretty decent examples of them changing the way they are handling both policy and implementation. Note that I am NOT SAYING they aren't still screwing things up, nor that they are prepared and capable of not screwing up with this adult content issue, as it is significantly more far-reaching and convoluted in its scope than all other things they have been doing combined. However, I think you're being a tad disingenuous saying that they are "pretty much unchanged" over the last four years.



Nope. I am saying that they should INVESTIGATE edge cases, when the circumstances warrant. I mean, how in the HELL are they supposed to know that they are "edge cases" unless they investigate? For edge cases that prove to be actual abuse, they SHOULD apply the policy liberally and fairly.

At any rate, regarding your billboard issues, the offer still stands; would you like a flash mob AR party to help get LL to notice the violation?

No, actually, we pressed Jack for TWO MONTHS about when we could start ARing landcutting and extortion plots, and he kept on saying things like this:

[2009/03/12 12:54] Jack Linden: Ah i see. If you guys aren't seeing significant change this time next month, you can come prod me. But you really should see some changes soon

In both cases, he never overtly indicated that ARs are "desired". However, they have been getting and responding to them anyway.

So, no, it is the same type of policy structure that they have been using.



Oh, so no one can see the list? Gonna kinda make it hard to use if it isn't ever for people to see. Kinda makes generating it pointless, to me.



Yes, that follows from your previous "definition"; if people were meant to "see" it, then it would be useful. Since it isn't meant for people to see, I guess they will have to just imagine what information is actually there and operate on that.

What you are saying makes so much sense now.



No, actually that is why I am asking the questions. You seem to be making out that you know an awful lot about it yourself, so I am asking of your (ostensibly) superior position of knowledge where they are getting it from. If it is not from Search, where ARE they getting it from?



Of course, but that's why I am asking you for enlightenment. Here's what I understand so far:

1) "monitoring" isn't "monitoring" if the info gleaned from the act of "monitoring" is not meant for people to "see".
2) "reports" aren't "reports" if they are not meant for people to "see".
3) They aren't getting the info that they are not meant to see from Search or from anything to do with Search.

Is that right so far?