This is the last LWN.net Weekly Edition for 2006; following our
longstanding tradition we will take the last week of the year off and
dedicate it to cleaning all of this year's unanswered mail out of our
inboxes. We wish you all a pleasant holiday season; LWN will be back on
its regular schedule on January 4.
Another LWN tradition is to review our predictions made at the beginning of
the year to see just how badly wrong your editor was this time around.
Those predictions were published in the January 4, 2006
edition, for those who wish to follow along from the source. Some of
the comments posted to the article can also be interesting to read with a
year's perspective. We'll not review every prediction made in that
article. Some of them are sufficiently obvious ("Perl 6 will not be
released," "the SCO case will drag on") or general ("the pace of kernel
development will not slow") that little review is called for. Some of the
others, however, offer some insights into how perspectives have changed
over the last year (or, perhaps, how blind your editor was back then).
The very first prediction made was that the GPLv3 process would dominate
the news. Your editor was not able to foresee, however, that the FSF would
take the license revision as an opportunity to attack DRM head-on. What
has happened over the last year, as evidenced by GPLv3 and in other places,
is that many in the community now think that we have enough weight to throw
around in support of goals beyond the simple creation of free software.
Whether the exercise of this weight
will lead to a more free society, or whether it will just make us more like
the entertainment industry (which also thinks it has plenty of weight to
use in pursuing power under copyright law) remains to be seen.
Some commenters doubted your editor's prediction that the non-free kernel
module issue would come to a head this year. But, over the course of this
year, a number of distributors swore off shipping such modules, those which
continue to embrace proprietary modules have taken a fair amount of
criticism, and the kernel developers seriously considered banning them
outright. Whether all that constitutes "coming to a head" can be debated,
but the fact remains: there is a great deal of resentment over proprietary kernel
modules and this issue will not go away anytime soon.
Your editor predicted the return of European software patents. There were
some stirrings over the year, but software patents have, for the most part,
laid low. It would be foolish to believe that they will do so forever,
With regard to desktop Linux, your editor's advice was to not expect
amazing advances, but that there would be steady progress. The movement of
3D technologies onto the Linux desktop may not qualify as an "amazing
advance," but they are a big step regardless; Linux need defer to no other
system in the eye candy department. A prediction that alternatives to
OpenOffice.org would gain prominence did not really come through - but it
is worth noting that the OLPC project has gone with a lightweight version
One of the more controversial predictions said that the Fedora Project
would have to make changes to maintain its position. Over the course of
the year, Fedora abandoned the "Fedora Foundation" idea, gave up
(belatedly) on Fedora Legacy, decided to lengthen its support period, and
merged the Core and Extras distributions. The project has picked up a new
energy, renewed its longstanding dedication to free software, and looks
well poised to move forward with a stronger community focus.
Predicting that a Debian release would happen on schedule is always a
daring thing to do. Things clearly did not work out that way, but
substantial progress has been made. Debian Etch might not be that
late, in the end. Predicting Emacs releases is equally risky, and
Emacs 22 did not come out this year - but a couple of pretest releases
Your editor thought that Novell would "get its act together and become a
truly successful Linux-based company." Oh well. That could yet happen,
but, after the events of 2006, few people would see it as a foregone
So what did your editor miss entirely? Big company moves were at the top
of the list. The idea that Novell would make a deal with Microsoft -
paying patent royalties in the process - was beyond your editor's
imagination at the time. Similarly, the notion that Oracle would try to
muscle into Linux support by repackaging Red Hat Enterprise Linux was a
surprise. Free software has reached such a level of importance that the
largest companies out there are paying attention.
Also missed was the open-sourcing of Java, though one could certainly
quibble that we have not actually seen the code yet. Perhaps your editor
should simply predict this event for 2007 and be dead-on. Seriously,
however, this event has been delayed for so long that many of us had
despaired of it ever happening. It does appear, however, that Jonathan
Schwartz has brought a new emphasis on free software to Sun's top position;
the planned release of Java under the GNU General Public License suggests
that he is serious.
In the end, the easiest prediction to make was that our community would
remain healthy, and that our software would continue to get better.
Despite our disagreements and our mistakes we are going from one strength
to the next. That helps make 2006 another pleasant year to look back on.
Comments (16 posted)
For the ninth year in a row, the editors at LWN.net have put together a timeline
highlighting the most important events of the last twelve months.
It has been an active and interesting year - just like the ones before.
The GPLv3 process was launched - and threatened to split our community over
differing views of freedom. Software patent issues came and went. The
Linux desktop went 3D. Large companies became more involved with Linux and
free software - and not everybody is pleased with the result. Distributors
reevaluated and reworked their dealings with the community. And,
while all this was happening, the community continued to produce great code
which made all of our systems better.
This is version 1.0 of the 2006 timeline.
If you find any errors or remaining major omissions, please send them
to us at email@example.com; please do
not post errors or omissions as comments until after we have had a chance
to address them.
The development of the LWN.net Linux Timeline was supported by LWN
subscribers; if you like what you see, please consider subscribing to LWN.
This year, we are pleased to announce the return of the one big page version as well.
For the historically minded, the timelines for the previous eight years
Comments (none posted)
When Larry Lessig
law" he was talking metaphorically. But for a virtual world,
constructed entirely out of bits, it is literally true: the laws regarding
what you can and cannot do there, both legally and even physically, are
inscribed in the lines of code that implement it. In this space, then, open
source has an added significance in that it not only lays bare the engines of
creation, but it potentially allows them to be hacked.
What some of the consequences of this openness might be was shown recently in
Second Life, when the
open source project
released a program called CopyBot. As its name suggests, this tool allowed copies
to be made of in-world objects - including the "avatars" that are used to
represent the residents of Second Life.
This was deeply problematic, since one of the attractions of Second Life is
that creators of digital content retain
unlike in most other virtual worlds. Many now make a good living from this
in-world activity selling virtual items, with
earning tens of thousands of dollars per year. However, CopyBot raised the
spectre of people replicating content for free, rendering digital objects
valueless, and undermining the entire Second Life economy.
The person leading the libsecondlife project is Jonathan Freedman. He took
over recently after John Hurliman, the previous lead, and still the main
contributor of code to the project, decided he didn't want to deal with the
public relations issues that CopyBot threw up. Freeman recalls: "he said to
me: 'I just want to code, I don't want to deal with this.'"
The libsecondlife project began six months ago, and was started by a group of
coders who "were interested in seeing a little more flexibility in what they
could do with Second Life," as Freedman explains. The idea was to create an
open source library that third parties could employ to create new Second Life
applications. To do that, the libsecondlife group started reverse-engineering
the Second Life protocols.
One by-product of this work was that they turned up security issues - "and
believe you me, they found quite a few," Freedman says - which they reported
to Linden Lab, the company behind Second Life. Partly as a result, "the way
the project had been run impressed Linden Lab, who were very happy with it,"
Freedman explains. "Back in the Second Life Community Convention in August,
they gave their unofficial
of the libsecondlife project."
And then along came the CopyBot incident.
"It was a debugging tool," Freedman says of CopyBot. "The developer was working on the
part of the Second Life protocol that was responsible for drawing avatars. He
needed a way to verify that the data was coming correctly: what better way to
verify that than just mirroring it back" down the connection to the system and
observing the result?
Freedman emphasizes that there were safeguards built into ensure that this
"mirroring" - copying of virtual objects - was kept within the terms of
service at the time. "You'd actually have to ask it before it would copy you,
and it would then give you a lengthy disclaimer explaining what was going on
so people could make sure that that was what they wanted. And generally people
were agreeing with that, and they'd be there for five or ten minutes dancing
There the story might have ended, were it not for the fact that CopyBot was
free software. "Anybody could get a copy and make use of it, and that's what
we saw happening: other people were modifying it to take out the disclaimer,
and generally shout stuff like 'I'm stealing your textures'" - the surface
elements of virtual objects.
As well as taunting victims in this way, a few of these "griefers" started
selling the modified, no-holds-barred version of CopyBot within Second Life.
Panic spread in some quarters of Second Life. Shop owners
hundreds of virtual stores, afraid that their inventory would be copied
endlessly and rendered worthless. But in practice, the
was minor, and the economy of Second Life continues to
not least because CopyBot itself had important limitations that were
consequences of the way Second Life operates.
Each "sim" or simulator of a portion of the virtual world in Second Life is
created on a server running Debian GNU/Linux, Apache, Squid and MySQL;
currently there are several thousand of these PC boxes. To allow for fast
response times, the virtual world is sent not as pixels or even as a mesh, but
as a series of 3D primitives - "prims". The Second Life client creates the
world by converting the stream of information about prims and their position
into a visual representation.
This means that the client has all the structural information about any object
visible to it; CopyBot works by taking that information, and replicating it.
However, in addition to the prims and the textures applied to them, more
complex objects add scripting to provide interactive behaviour that endows
Second Life with much of its richness. These scripts are run server-side, and
are not passed to the client, so CopyBot is unable to intercept them.
Nonetheless, the residents of Second Life who made money from their virtual
creations were understandably perturbed by the appearance of a piece of
software with the provocative name of CopyBot - "in retrospect it probably
could have been named something else," Freedman concedes.
At a November meeting held in-world, Second Life's creator and CEO, Philip
that nothing could be done about CopyBot using technical means: Second Life's
client-server architecture implied that CopyBot was not just possible but in
some sense inevitable. But he did promise other measures, including more
metadata, such as attribution and creation time-stamps, for virtual objects.
Since these would be stored server-side, and hence immutable, they would
provide clear proof of whether an object had been copied. To give this
approach some teeth, Linden Lab made
that anyone who used CopyBot or something similar in a malicious
manner faced the
prospect of expulsion from Second Life.
with Rosedale's response, and also see the CopyBot incident as part of a
deeper malaise involving cynical hackers exploiting loopholes in the Second
Life code to grief other residents. They accuse Linden Lab of a certain
complicity because of its encouragement of the external libsecondlife project.
Perhaps that encouragement is not so surprising given Linden Lab's stated
intention [PDF - look at final slides]
to make elements of Second Life open source. "Without speaking to specific
timing or plans - and we've thought and are thinking lots and lots where there
might be exceptions to this - it seems like the best way to allow [Second
Life] to become reliable and scalable and grow," Rosedale said
on the subject of opening up the code. "We've got a lot of smart people here
thinking about that." It's obviously useful to have smart people thinking
about it on the outside too - provided things don't get out of hand.
Freedman has instituted one important change in the libsecondlife project to
try to ensure that another CopyBot does not happen. "Previously, the way the
libsecondlife source tree was done was basically anybody who wants an account
can have one. That's the first thing I changed: just the
developers can have the accounts."
Freedman also has some clear-cut goals for the project, which will be
releasing all its code under the BSD license. "Short-term, the aim is to have
a workable third-party library that other people can make use of to interface
with Second Life. I believe that by the middle to end of December we'll have a
fairly decent third-party viewer that's comparable to the Second Life [client]
application. Longer term, ideally we'd like to see a completely open
implementation of Second Life, from the client, to the sims, to the assets -
Freedman believes "the use of open standards, if not open source, will go a
long way in the propagation of Second Life as an actual platform." This seems
to explain Linden Lab's enthusiasm for libsecondlife and patience with things
like CopyBot. At stake is the chance to help create the next online platform -
the 3D Web, sometimes known as Web 3.D.
Opening up the platform will also take some of the strain off Linden Lab:
currently, Second Life is growing at an unsustainable rate, with over a
million new members joining in the last couple of months. If users could host
their own virtual land, then Second Life could scale more gracefully. Beyond
that, open protocols would allow distinct but
virtual worlds to be created. The technical aspects of this are the easy part;
more difficult are working out social and economic issues like making
reputation and money portable between those worlds, and legal ones - as the
CopyBot episode made all-too clear.
Glyn Moody writes about open source and virtual worlds at
Comments (47 posted)
Page editor: Jonathan Corbet
Next page: Security>>