A brief history of content-ripping

This “little” post has come about as the result of a suggestion from regular reader, Peter Stindberg, which followed the concerns I raised about the “Bye Bye Copybot” prim being circulated by members of the Emerald team, and lauded by  some content creators and others, despite its potential as a ToS violator.

Before I get to the nitty-gritty, I will point out that while what follows is a genuine attempt to timeline events and interrelations of events as accurately as possible, the degree of paranoia and misinformation surrounding content ripping means that a) there is a possibility I may fail to mention some events ; b) some may view things differently (and may themselves not necessarily be correct); c) I’ve confused issues (although I’ve verified as much as I remember over the years with various sources elsewhere) – in which case polite corrections welcomed!

libsecondlife

Up until 2006, Second Life had largely been a closed universe: the code for both the server and client-side software was developed purely by Linden Lab. There had been issues around copying content – tools like GL Intercept, which enabled the likes of avatars to be copied, and basic tools and scripts that enabled textures to be pulled from the local cache. However, these tended to be somewhat obscure as far as the populace of SL at large were concerned, which somewhat mitigated their impact (but didn’t excuse their use).

Then something happened: libesecondlife was created.

libsecondlife commenced, with Linden Lab’s blessing, as group of Second Life residents attempting to back-engineer an open-source version of the Viewer (client-side software).  Today, to avoid copyright issues in relation to the Second Life name the group is now called lib.openmetaverse.org.

CopyBot

As a part of this work, an automated tool was developed – CopyBot that could be used to replicate avatars – bots, and be used as a debugging tool. The original CopyBot required that that target user actively give permission to be copied, and issued a disclaimer (in the form of a drop-down) prior to the copying taking place, specifying the fact that ownership / permissions pertaining to anything worn by the avatar would be lost in the copying process.

The software itself (written in C#) worked by intercepting the communications between the Viewer and server and replicating the information relating to objects (prims, textures). In doing so, and due the the way in which Viewer / server communications had been coded by LL, the copy process would lose the metadata relating to the original creator of the object and all permissions set against it – there was simply no way of including this information in the raw copy process as written in C#.

The original version of CopyBot was published as a part of the libsecondlife library of tools, where it became relatively easy for someone to remove the code asking for the target’s permission to be copied and the disclaimer drop-down, and thus use the code to copy virtually anything in Second Life – with the exception of scripts, gestures and animation (although later iterations of CopyBot could apparently grab animations and gestures from avatars) – completely surreptitiously, giving birth to CopyBot as we know it today.

What made CopyBot different to earlier attempts to rip content was its relative ease-of-use, it’s availability and – inevitably – the notoriety it quickly gained as a result of being made “public”.

From the start, the libsecondlife group were fairly unconcerned by the risk CopyBot presented to content creators, demonstrating a “so what?” attitude, supported and repeated by other techies posing as “journalists” as the news broke. Indeed, in one such interview, libsecondlife’s Admin, Babba Yamamoto intimated the issue of metadata loss could have been overcome if the CopyBot code had been re-written in XML (this is in fact how tools such as Second Inventory and Viewers such as Meerkat and Emerald “legally” export content)  – but no-one saw the point in doing so, since the “flaw” that lost the metadata lay with the way LL had originally coded Viewer / server communications, so those in the libsecondlife group felt justified in deflecting anger directed at them by pointing the finger at LL.

libsecondlife did eventually pull the CopyBot branch from their open source library as the wave of outrage reached the level of a tsunami – but again (and to mix metaphors) their action was that of not only shutting the stable door after the horse had bolted – but having ensured the horse had an open-ended ticket to any destination of its own choosing as it left the stable.

Protests and Response

The “revised” CopyBot code quickly started showing up as being for sale both in-world and on SL Exchange and immediately generated uproar as a result. Protests were held, the forums were flooded and people were angry: stores were closed; sims locked down, and calls were made to boycott Second Life.

While the libsecondlife group’s reaction remained pretty much, “so what?”, Linden Lab’s initial response to the protests could best be described as lukewarm, with Robin Linden repeating the assertion that content copying is “not necessarily” theft – a meme initially rolled out by Cory Linden – himself an active supporter of libsecondlife. While the meme is technically true (it should be pointed out that Second Inventory, for example, effectively uses CopyBot-style coding with ownership and permission checking in place, to export objects from Second Life, for example), it was also was somewhat disingenuous to raise it in the context of content ripping.

Many were angered by this reaction from Linden Lab, which gave rise to a further round of protests. Some of these made headline news in the likes of Business Week and  News.com – prompting Linden Lab to take something of a more affirmative stance, revising their policy to make it clear the use of CopyBot and similar tools would not be tolerated. While the move was in some ways welcomed, it was also felt that it was the risk of poor publicity, rather than a desire to help reduce the risk of Copybotting that prompted LL into “action”.

Nor was the revised policy that successful. By referring to the use of such tools as CopyBot, the policy implied that the sale of such tools in-world was still OK – and so people kept right on selling it at up to L$1500 a pop, quite prepared to face the wrath of residents while the Lab again kept quietly to the sidelines.

At the same time as calls for tougher action against CopyBot and it users continued, so to did the counter-argument “that nothing can be done” to stop the matter (an argument still heard today) gained strength among techies. People pointed to the code behind web pages being viewable and therefore copyable; people pointed to the existence of tools such as GLIntercept that could copy avatars, people raised the issue of ripping MP3s over the Internet as reasons why CopyBot was not only “inevitable”, but should be more-or-less accepted.

While such statements are broadly true, they in no way justify the theft of Second Life content – or any theft for that matter. Rather, they deflect discussion from the core issue that Second Life is promoted as a platform of commerce, and as such those encouraged to take up business opportunities on the platform should be offered a degree of protection that appeared to be somewhat lacking on LL’s part.

Promises and Tools

Another reason people perceived Linden Lab as having little concern over the matter was the time taken to develop practical tools that could help in the identification of potentially ripped goods. Conversations around such tools commenced in November 2006, with Robin Linden’s above-mentioned post. However, it was not until April 2008 that the first really useful tool –  the Object Inspector – finally made its debut;  18 months after the initial furore.

Now, there could be perfectly legitimate reasons as to why such a tool took so long to develop and deploy. However, during the 18 months it was in development, Linden Lab was largely silent on the matter of content ripping, giving rise to the perception that they “weren’t interested” in dealing with the issue. Right or wrong, this perception was further reinforced by the fact it was not until August 2009 – nigh on three years after the original protests – that the Lab saw fit to outline updates to their IP Complaints Process, as outlined in their Content Management Roadmap. Even then, insult appeared to be added to injury in that elements of the Roadmap appeared less concerned with the worries of current residents as they did in providing the perception of content protection for “upcoming” users of the Second Life Enterprise product.

In the meantime, CopyBot development continued among known hacker groups – such as the Patriotic Nigaras – who worked to make the tool more “user-friendly” and added further capabilities to it. While the sale of the tool was banned from both in-world and on the likes SL Exchange / XStreetSL, CopyBot does continue to be available through various torrent sites – although I understand (but have absolutely no proof) that many of the advertised CopyBot downloads are, in an ironic twist, themselves riddled with viruses / Trojans.

Viewer Threats

Throughout 2007 and 2008, CopyBot remained an issue. Exactly how widespread its use was was difficult to assess: paranoia meant that many reports of copying came via “friends of friends of friends” rather than first-hand exposure, and the storm was further whipped up by merchants selling “anti-Copybot protection” tools that were little more than scripted placebos. While such tools did little or nothing to stop Copybotting, their widespread proliferation  in stores across the grid reinforced the perception that Copybotting was epedemic in proportions.

Then the landscape started to change. In 2007, the first fully-functional third-party Viewers began to appear. Over a period of several months, a crop of Viewers showed up that offered people a genuine alternative to the “official” Viewer, which was regarded as poorly-written and crash-prone. What was more, these Viewers not only offered improved stability, they also tended to offer features that users has been requesting from Linden Lab without any success. As the popularity of these Viewers increased, so did the likelihood that the Viewer code would be maliciously hacked.

This likelihood became a reality in 2009, when the first of the “CopyBot Viewers” appeared in-world. This did raise content theft to a new level: now it was possible for anyone to grab content (with the exception of scripts) simply by using a modified Viewer – no other tools or add-ins required.

As with CopyBot itself, Linden Lab were initially slow to respond, despite the renewed outcry the appearance of these “hacked” Viewers caused. It was not until February 2010 that their Third Party Viewer (TPV) Policy first appeared.

BuilderBot

In mid-2009 the matter almost took another very nasty turn when Jim Himoff of Rezzable suggested an in-house tool created by his company would be made available to the open source community. This tool was Builderbot. Based on the CopyBot code, Builderbot enabled entire sims to be “backed up”: land, buildings, content, textures – all in a single pass.

Like CopyBot, Builderbot was initially developed with a genuine function in mind: Rezzable had invested heavily in Second Life in terms of sim development and commissioning custom builds they have not only paid for, but have purchased the rights to as well. When they opted to make a move from Second Life to their own OS Grid, they obviously wanted to take their investment with them: hence Builderbot.

However, Himoff’s announcement of issuing Builderbot to the open source community as an unrestrained tool was alarming, and his very public claims that it presented “nothing new” as it was based on “CopyBot” and the “CopyBot was already out there” was an utterly disingenuous excuse. Given the history of CopyBot, any release of an unrestrained version of Builderbot would have been worse than mischief making – it would have been as malicious as pulling the pin of a hand grenade and tossing it into a crowded room.

Fortunately, such was the outcry over the announcement that Rezzable quickly backpedalled away from their stated intent (assuming said intent wasn’t a stunt aimed at raising Rezzable’s visibility and that of their new OS Grid offering), opting instead to develop a version of Builderbot that would include ownership and permissions protection. So far as I’m aware, no version of Builderbot has to date been released.

The Present

Most recently, Linden Lab again caused some consternation with the release of Viewer 2.0 – which had the invaluable Object Inspector completely removed. Why this was done is anyone’s guess, but it lead to some consternation on the part of residents using it – so much so that a JIRA was raised and LL reintroduced it with the first Viewer 2.0 update –  although one has to ask why it was removed in the first place.

We’re still awaiting further updates around the Content Management Roadmap – again nothing has been heard of on this subject since the August 2009 blog post, giving the (potentially wrong) impression that it has fallen of LL’s radar in the rush to get Viewer 2.0 and all things “Enterprise”-related out the door.

Currently, the subject of content ripping remains contentious. Not because it isn’t happening – it is – witness the recent XSL sale of ripped dances and Ishy Wingtips’ highly-literate and thought-provoking flog posting on the subject as it affects the Teen Grid –  but rather because the accepted perception is that copy ripping has reached pandemic levels in Second Life. This perception has resulted in a lot of misinformation to enter circulation – some of it through simple misunderstandings, some of it to deliberately derail attempts to halt the further spread of the problem.

The waters have also been muddied by the subject of “copybotting” being used to promote other agendas. Towards the end of 2009, for example, a number of high-profile content creators used the subject of copybotting to float the idea that only a selected “elite” (my term) of content creators should be allowed to operate in Second Life. Among other things, they suggested the criteria by which such creators should be selected should be related to in-world turnover – an idea that found its way into the Content Management Roadmap thus: We are starting the process of planning a content seller program, and we would like your input on possible program criteria. At a minimum, participation in the program will require that the selling Resident…..3. meet a minimum threshold for content transactions.

Quite how limiting the number of content creators to a “selected few” would stop content ripping (given their own content would clearly become the target) has never been fully explained – but the fact that their views, issued under the guise of concern about content ripping, found their way into the draft CM Roadmap is disturbing.

Beyond this, and in part due to the apparent lack of concern from Linden Lab on the matter, in-world tools to “combat” Copybotting have appeared over the years. How effective any one of these tools is, I cannot honestly say. Some appear utterly useless and smack of cynical attempts to cash-in on people’s fears. More recent tools have appeared that seem to offer a degree of protection against the use of “hacked” Viewers, but even these are subject to considerable controversy, inasmuch as a) they have been developed by former content rippers (allowing their potential effectiveness to be undermined using the question of trust), and b) the creators have been so secretive around the tools, entire rumour mills have been created around them- and not always positively.

And so we come bang up-to-date with things, and the post that initiated Peter’s suggestion that I try to summarise (!) things relating to Copybot, etc.

I think I’ve covered all the bases and key events. If I’ve missed anything significant, I apologise, and will attempt to correct any omissions / inaccuracies that are pointed out to me.

Further information on CopyBot and the furore around content ripping can be read at:

Note: Revised Mar 26, 2010 01:45 BST to better reflect the situation prior to the advent of CopyBot. With thanks to Tateru Nino for both prodding me in the right direction and for giving further information.

4 thoughts on “A brief history of content-ripping

  1. You seem to have missed the content rippers that came before libsecondlife and that used the official viewer. They started in about 2004, I believe.

    A good effort though!

    Like

    1. Tateru. Many thanks. I was initially involved in SL from early 06 for a while (different avatar name). At that time I was not involved in content creation; as such many of the nuances of SL passed me by. I’d be grateful of further information, so I can put things fully into perspective, timeline-wise. I’m aware of the likes of GL Intercept – is there anything else?

      Like

  2. Mostly (as far as I’m aware) the tools were GL-intercept (and other debugging and interception libraries of that stripe – last I looked there were a couple dozen of those) or simply yanking textures right out of the viewer cache and converting them to another format (the images are split into pieces, but it’s simple enough to put them together, and there were some simple tools and scripts circulating for that).

    There was also the use of LSL to export prim geometry, though you had to have *some* permissions to make that work. Copied objects could then be reimported by copy/paste in chat, via XML/RPC or through email to scripted objects.

    Like

Comments are closed.