A thread every content creator must see.

The underlying method: it's impossible to display stuff on your screen without letting you copy it. No change to texture baking will stop it, though the coming move to owner-only baking will make it a bit harder. It doesn't stop people from ripping hair, shoes, or buying skins and ripping them or, even with server side baking, buying clothes and wearing chromakeyed skins to rip the clothes layers. There is no technical solution. The only possible responses beyond what they've already got in the pipe are:

* Increase enforcement.

* Decrease the incentive to rip stuff.

Not. What. I. Suggested.

Not. What. I. Suggested.

Not. What. I. Suggested.

I don't sell clothes and I'm not butthurt by the advent of CryoLife, NielLife, ThugLife, or RipoffLife (I'm the guy saying "don't panic", remember). This would have no effect on me. It wouldn't make customers "walking billboards" any more than they are now... whether you right-click on a product and take one step or 10 steps to buy it, you're still just as much a "walking billboard". Or do you right-click on billboards and buy stuff from them in RL?

Oh, come on, Kitty, that's not fair. I raised the idea and Argent mentioned that he'd thought of it long before. You know that Argent gives away more stuff than he sells, and I can't even give my stuff away. I hardly think shilling for sales is on either of our priority lists.

In any case, I agree it seems a little crassly commercial, but it's not really different in kind from what's there now.

Lots of resident discussion in numerous places. The sound of indifferent silence from Linden Lab on this topic is deafening and very telling.

Could it be that behind the curtain Linden Lab is rooting for the looting?

The deafening silence is nothing new.

In addition, they are most likely under "litigation gag order" due to Stroker et al's suit, so they are likely even more quiet about it.

Ick. It never ceases to amaze me the depths to which people will sink just for what amounts to a few bucks. These people are parasites in the truest sense of the word: They invade a host, feed off of it until it hobbles over and dies, then move on to another host.

I totally and completely blame the hip-hop lifestyle and all of the people who embrace it for elevating the term "thug" to mean anything more than the lowest form of animal life in existence.

Anyway, enough ranting. I think I may have an idea that will stop all this crap once and for all, WITHOUT jeopardizing the spirit of open-source. The idea is this: Each client has to authenticate itself to the server. It will require that third-party viewer developers submit their viewers to LL for screening and approval, simply so that LL can verify the viewer won't bring harm to the grid, and so a hash can be generated and registered. Each approved viewer will then include one external closed-source module to be installed along with the client. The module's sole function will be to generate a hash value for the viewer binary and respond to authentication requests from the grid's authentication server.

When a viewer connects to the grid, the authentication server sends a randomly-generated hash along with a hash request. The closed-source module will receive the request, generate a hash value of the viewer binary currently in execution, combines it with the randomly-generated hash it received from the server, and sends the combined hash as a reply. The authentication server will then reverse the combine operation, check the viewer hash against its database of known viewers, and grant access if it matches.

I'm sure there's something I'm probably not considering, but what do you all think?

This mechanism is already widely used in the video game industry, and everyone I know who plays these games for very long is using SOME kind of so-called cheating tool, if only something to automate idiot aspects of the game like fish-mining. These tools have to trick a module like that to work. They're always available, even right after an upgrade.

All this will do is make it harder for open source viewers to be developed (when you're developing the viewer, you have to recompile it over and over again, which changes the hash), while the crooked viewers will just do a two-stage setup... running the real viewer under a debugger to find out the correct hash and then patching the module to always think it's got the right hash. This kind of thing is completely automatable, and common.

Video game cheats do not function via a modified client. Video game cheats exploit the weakness of game parameter values stored in memory. Video game creators try to obfuscate these locations but it is trivial to run patterns and detect the patterns and have a new cheat out in minutes. Punk Buster scans the drive for external modules so the serious cheat systems mask themselves to something random and different on each run thereby rendering punk buster pointless for the serious cheats (sold cheats).

Second Life is completely different having become open source. However even with the hash value trick it would not stop the slproxy method used before SL went open source.

The solution is to take out the humans involved, via legal means of course, so they will no longer feel they nothing to fear from attempting to destroy a billion dollar business. Once the first few fall the rest will flee to other places to conduct sociopath attack operations and "cause grief".

LL has to eradicate the source not the symptom. People who need Cryolife and ThugLyfe to steal are not going to be writing a theft client in the next decade. Remove the theft client programmers and the problem is mitigated. But LL will not do this and why is the question that needs to be answered.

That wasn't the point I was trying to make though .

Most of the concern specific to the new viewer seems to be about "casual copying" things people wouldn't already own which pretty much dissapears for body parts/clothing if LL stops sending clothing layers around to anyone except the person wearing them (except for eyes I guess since those are their own baked texture).

It's also one of the very few things LL already agreed to and wouldn't really impact anyone (except those on pre-1.23 viewers).

With that in place and the exploit fixed LL didn't have to actually do anything about the viewer (which would run into the standard - if true - "it's just not possible" but it will still negatively impact it a bit (from the point of view of the copier) which is really what everyone's after to begin with.

That's what I meant with don't focus on the tool but on what/how it's doing it.

Some of the simpler ones don't. The ones that change the behavior of the game do. Whether they modify the client on disk or in memory (there are examples of both) they are still actively searched for by modules exactly like the one being proposed.

I think we're basically in agreement that this should be treated from the enforcement side.

I do appreciate your presence on the forums, you make me seem such a moderate.

Yeh, but that doesn't need server side baking, which was what I was responding to.

And it still wouldn't protect hair, shoes, jewelry, prim avatars, vehicles, ...

Some do. Some modify the game data and metadata. Some modify the executables on disk. Some use hooks to intercept system calls. Some act as packet proxies. Some use virtualization. It depends entirely on the nature of the cheat, the game, and any/all detection/prevention methods used to thwart said efforts.



It isn't "completely different". Open source is nothing more than an ADDITIONAL vector. LibSL most certain predated SL's Open source, and would likely seen a great deal more development focus if LL hadn't opened the source.



Which will never be successful for any number of reasons.



No, but someone, somewhere will always see a need to do it, and will provide it. The more LL tries to chase it with the law, the more out of reach it will become to them; however, it won't be out of reach to casual users.

It is a waste of time and effort best spent elsewhere.

Case in point: WoWGlider.

Talarus your constant leftist west coast geek style support for content theft is not unnoticed. People need to make a mental note that you are a cheerleader for content theft with your constant belittling of any discussion on the topic.

The fact is if the PN were arrested and thrown in prison to be meat bots in the showers then others would stop cold and no longer be interested in writing theft systems for Second Life. They would simply leave if they were so mentally ill that they needed to get sexual pleasure from causing grief and stealing.

What good is a system if everyone quits?

Ann: that's a nice image, but I don't see it ever happening.

Sorry bu I didn't find WoWglider enhance my enjoyment of WoW in the slightest, it made it worse watching bots just farm the crap out of areas ignoring any basic ettique, farming nodes while you were still fighting the critters guarding it.
It's cheats tool, Idon't know what all the rush is to get a character to retirement ASAP anyway.

Maybe. If citizens began reporting the situation to the authorities in a manner they would listen to then LL may not have any choice but to comply with the ensuing investigation.

People keep asking for solutions. I present the only one that will work. Naturally people with something (like freedom) to lose will vehemently argue against the only solution that will work.

There is one other solution I alluded to earlier. Pattern recognition + license metadata. When you release a product you have it run through a pattern recognition system and input your licensing data. Each time you sell the item the buyer is registered as a licensed owner. Anytime someone imports or acquires something that in whole matches a registered pattern it is blacklisted and LL alerted to review how the content came into existence. If the logs show probable cause then the account is deleted along with everything that account ever created being blacklisted as suspect. The problem is one of resources in addition to the ever present who watches the watchmen paradox.

So you either have a law enforcement approach that imprisons ripper client authors (for whatever charges will stick and there is always something. Ask Max Hardcore about his 46 months) or heavy coded systems that are effectively doomsday devices. I would prefer to send anyone willing to participate in writing theft systems to prison where the really bad guys can use them for entertainment. Then, if they survive, they can live at underpass motel for the rest of their lives. All because it was "for the lulz".

This is sort of a follow-up to my naive post across the street about server-side baking of avatar textures. The subject seems of more interest on this forum. It's probably hopeless, but I'm going to blunder ahead anyway.

There are (at least) two big objections to server-side baking. First, it's not scalable. I'm surprised by that because the operation as I understand it is very simple--it's not 3D rendering or anything--so I didn't even think it used the graphics card GPU. But how many times do people change clothes, compared to all the other graphics operations that are done for rendering an avatar on a viewer? If baking really does rely on a GPU, surely a single, extremely primitive chipset on the server could satisfy the costuming of hundreds of avatars, couldn't it?

That's assuming the baking operation is as simple as I think it is. Maybe it shouldn't be, and that gets to the second big objection: the ability to "unbake" a layer just by wearing it alone on a chromakeyed avatar. But what if baking were somewhat more complex, and included a non-invertible transformation on the image?

An example of such a transformation is distortion between UV maps. Suppose the individual clothing layers mapped to the current avatar mesh, but there was a new avatar mesh with a different UV map. It would be possible to distort between those UV maps, but it would introduce blurring. Distorting back from the avatar map to the clothing map (ripping a clothing layer from an avatar texture) would introduce yet more blurring, which would only get worse when the doubly-distorted image is distorted yet again to the avatar map (wearing ripped clothing).

Is this going anywhere, or should I just give it up already?

Artists are already ticked off enough by the distortions in the current UV maps.

I think you're up against the fundamental DRM paradox.

That code gives me a headache so I can't say all that much useful about it .

Compositing the textures is definitely done on the video card though (or more accurately it's done using OpenGL which may or may end up using the video card depending on the card and driver).

And changing clothes probably does generate a whole lot of overhead: if I switch from one dress to another by removing what I'm currently wearing and then wearing the new dress then that's at least 8 partial rebakes (removing the jacket invalidates upper and lower body; removing the shirt invalidates upper, removing glitch pants invalidates lower and then the same in reverse).

Not to mention that you get a local rebake every time you move the length sliders or change a tint in the appearance editor. Downloading a new baked texture every time you twitched the mouse would probably be a bit laggy.

Just opening and closing the appearance editor causes a rebake. (at least it appears it does) It is a technique used to make yourself rez in sims you enter. on landing move out of the way and then edit appearance a couple of times to force your shape to load in that sim and then rebake in that sim. Then others in that sim are not forced to download your shape & bake from wherever you last baked. A spotty operation that apparently fails routinely in a sim with more than 20 avatars. So having the local sim contain your appearance makes sure you might not be a fart or gray bot.

Unless they recently changed this. From what I have seen things still seem to function the same way.

People are more likely to steal hair and shoes anyway and no amount of shake-n-bake is going to alter that nor slow the thieves down. Besides once the skins get harder to steal they will find ways to get them and pass them around to friends who pass them around. Again it is a social issue that must be cured with merciless draconian enforcement that drives off the teenagers who happen to be the problem and are not even supposed to be there. Guess LL will have to find a way to ID people by DNA samplers eh? No. Not ready for a forehead bar code yet.

LL needs to identify miscreant coders and cause them to not want to code theft clients. Doesn't LL need to hire programmers anyway? Laugh all you want but it is a damned effective way of eliminating such issues rofl.

The quicker they can level it, the quicker they can return it to the original owner and start working on the next $150.00 power level.

There's only one way for them to do this, which is to disallow 3rd party viewers, and include a crypto-secure challenge-response on loging so that only LL-distributed viewers are allowed to log in.

Other than that, there's no way to prevent content copying.

Given that content copying is technically possible, the only mechanism to fight it is to be vigorous in pursuing claims against abuses. Of course, keep in mind that many cases will be bogus, so diligence is necessary.

Of course, as I hope we already agree, they need to fix the bug that allows copying of content in inventory regardless of object permissions.

I don't know what else they could do.

The "buy from XStreet" pie menu item is slick, but has problems.

Would not do any good as all they would have to do is channel the viewer through a proxy with a nice UI.

Take a look at PAR, which was LGG's first project before GreenLife. All of these functions have been integerated into the GreenLife viewer now.:

http://code.google.com/p/par/

Good point: communications would have to be encyphered, as well (not all, but much of it). Regardless, I doubt LL will bar 3rd party viewers.