GUADEC: Luis Villa points GNOME at the web
Longtime GNOME developer and community member Luis Villa kicked off the GNOME users' and
developers' European conference (GUADEC) with a challenge to the
project to "embrace the web
" as a way for the project to
remain relevant. The web has won the battle to produce a "robust,
libre platform
" over various desktop efforts like GNOME, but there
is still time for the project to find a seat at that table. It is a
"big scary step
" to take, Villa said, but one that he thinks
is ultimately the right direction for the project.
While he is currently working for Mozilla, which might have colored his
thinking some, Villa certainly disclaimed (in true lawyerly fashion) that
he was representing anyone's views but his own. He was taking vacation
time to attend the conference and wore a shirt from a company (Ximian) that
"
no one can be pissed at any more
". He was there because
"I love GNOME
", he said.
Villa was speaking from "the other side
", referring back to a
talk he gave at GUADEC in 2006 when he was "vanishing into the bowels
of law school
" and told the project members that he would see them
on the other side. That historical perspective was a major element of
Villa's talk; one theme revolved around a picture of a party on a Paris
boat at the first GUADEC in 2000. He considered what one would tell the
folks in that picture about the progress that has been made in the ten
years since.
Today there is a free and open platform that runs on all PCs and laptops, but which also runs on phones and televisions, a fact which would likely surprise the crowd from 2000. Most people using that platform also use Linux every day; the licensing of the platform is generally LGPL or more permissive. Even Microsoft has an implementation. There are some 400 million users. High school kids are learning to program for this platform partially by using a "View Source" button. Unfortunately Villa would have to tell those folks that this platform isn't GNOME, it is, instead, the web.
So the question is: what should GNOME do about that? Villa described
"one possible answer
", which is for GNOME to join forces with
the web development community and bring its strengths, in terms of
technical ability, culture, user focus, and hacker mentality, to that
party. GNOME should figure out how to deliver the best combination of
desktop and web applications to users.
Basically, the web won because it "co-opted our message
", he
said. He pointed to the famous Gandhi quote
("First they ignore you ...") but noted that things don't always work out
that way. "Sometimes your ideas win without you
", he said.
But, the web didn't win because it is perfect for either developers or users. There are problems with proprietary applications as well as control and privacy issues. It delivers sophisticated, powerful applications, though, which are run by someone else, freeing users from that burden. It's not a fad, and not going away, as it will only get better, he said. He also said that he had pointed the audience to an EtherPad site as a way to send questions, rather than to a Gobby instance, because he could be sure that all the attendees had web browsers while many would not have Gobby installed.
He noted that Apple and others brag about a thousand new "apps" this week,
but said that there are a thousand new web applications every hour.
Developers have already embraced the web in a big way; GNOME needs to
get on board. It is extremely easy to develop a web application by putting
some HTML and JavaScript up on a site somewhere; GNOME needs to be
thinking about making development that easy for the GNOME platform.
His suggestion was to start with "baby steps
" by
reimplementing the web's ideas for the desktop.
The web should be treated as a first-class object and various desktop applications should integrate with web services, he said. He pointed to the GNOME background image chooser which now allows picking images from Flickr or other web photo sites as an example. Though he noted that Zeitgeist hadn't made the cut for GNOME 3.0, he saw that as a step in the right direction because it treats the web as just another object.
Beyond that, the project should be thinking about even bolder strategies
that would not just copy what the web is doing. It will be bigger and
harder step, but he suggested that GNOME start writing code for the
browsers to provide any needed functionality. "Bring our ideas,
bring our code
" to fix areas that don't work for GNOME. As a
concrete proposal, he thought the Desktop Summit being planned for next
year (combining GUADEC and KDE's Akademy conference) should be renamed to
the "Free User Software Summit" and include browser developers from Mozilla
and Google.
Further out, GNOME should "burn the boats
" by writing all of
its applications in HTML and JavaScript first. Only when that doesn't
work should there be a fall back to GTK. According to Villa, GNOME needs
to start thinking that way because "that's how our users and
developers are thinking
". Instead of pointing developers at C and GTK or
PyGTK, GNOME should provide a first-class toolkit for HTML and JavaScript. It
should also be made easier to run the same code on the desktop or the web, he
said.
He allowed as to how this would be a major upheaval;
"I told you this would be hard
".
While it is going to require lots of new code, and potentially
abandoning lots of old code, it is still an embodiment of "our old
culture
".
Bringing that culture of freedom and user-focus to the web is Villa's
prescription.
For his part, Villa "welcomes skepticism
". Maybe folks think
the web isn't free enough or they hate JavaScript, but if so, they need a
counter-narrative: "Maybe my answer isn't right, but what is?
"
Maybe there are those that think the web is a fad, but they need an
argument to back that up.
He is optimistic about the future because of the people that make up
GNOME. "We are the right people
" to do this job, but need the
right code.
The clear indication from the talk is that he's convinced that the GNOME
project's current
direction isn't right and that a radical shift in focus is needed.
"Whether you agree or disagree or think
I'm crazy
", the challenge is to identify the right direction and
"go out and do it
". Villa has presented his idea of what that
direction should be, and he clearly thinks others should do the same.
| Index entries for this article | |
|---|---|
| Conference | GUADEC/2010 |
Posted Jul 29, 2010 1:02 UTC (Thu)
by karim (subscriber, #114)
[Link]
So, for better or worse, there will never be a "year of the Linux desktop." But we are most definitely, and will continue to be for the foreseeable future, in the time of the Linux server. We've also most certainly entered in the decade(s?) of the Linux mobile. But both in the case of server and mobile, FLOSS is taken for granted as a building block for far more richer applications. Android, for instance, bares little to no resemblances to a Linux desktop, yet it runs the Linux kernel.
I personally feel that somehow becoming webish is so far away from what Gnome does today that it's like trying to make an elephant fly. However, I would see how Gnome could revolutionize the mobile world. The fact of the matter is that Android is only open in the marketing sense. And the other Linux-based efforts are *way* behind Android. Gnome could most definitely be an attractive platform for pads and, with some zeal, for phones. But that would require rethinking some of Gnome's core paradigms. One thing that Gnome has in its favor, though, is that it's already started with a minimalistic mindset (recall Linus' take on Gnome's interface from a while back ...)
I know Gnome has already tried to make itself relevant in the mobile world, but that really went nowhere ... Now is a good time.
Posted Jul 29, 2010 1:50 UTC (Thu)
by skvidal (guest, #3094)
[Link]
He's absolutely correct and we have a heckuva a headstart. Just choose a infrastructure and write to it.
Posted Jul 29, 2010 2:31 UTC (Thu)
by aigarius (guest, #7329)
[Link] (9 responses)
But even worse is the JavaScript. It is a slow and ugly language that most Web developers I know despise with a passion. They'd rather code in ActionScript and have their app be locked into a Flash/Flex thing than code in JavaScript sometimes, they'd rather code in Java and have that translated to JavaScript.
While there might be some merit in replacing Glade with HTML+CSS templating, it would be much wiser to call Python the main programming language of Gnome and not even try to use JavaScript.
Posted Jul 29, 2010 4:11 UTC (Thu)
by jtc (guest, #6246)
[Link]
IMO, it would be very nice to see Linux-based (and FOSS-based) solutions that implement this new paradigm appear well before Microsoft has a chance to do anything about it. (Once the desktop [in its current form] becomes obsolete, my hope is that MS will, as well, become obsolete.)
In 5 to 10 years, people will be using systems that are much more integrated than they are today. Joe will pause his TV program (which he is watching either on his "TV", or his "computer monitor", or his "phone", or ...) to check his email, or do some research on the web (on a topic raised by the TV program he's watching), or send a message (whether it be a tweet, an email, an IM, a something-not-invented yet), or check his phone messages or etc. He'll choose a movie while at his computer, pause it to take a break, and finish watching it on his "TV". We'll have a highly-integrated, internet/web/cloud/...-based system that allows us to do many different things from any location and, hopefully, the monopolists will not have figured out how to leverage this system in order to force everyone to use their technology to access it.
Posted Jul 29, 2010 5:21 UTC (Thu)
by elanthis (guest, #6227)
[Link] (5 responses)
Your developer friends that would rather code in ActionScript (which is just JavaScript + static typing, seriously) or Java (or the subset there-of that certain tools translate to JavaScript) aren't the people to take advice on languages from. Those are the kinds of people who could only learn how to code by attending training seminars and community college classes (which are almost entirely all ActionScript and Java based) and could never manage to translate their cookie-cutter cut-n-paste programming skills to another language without having the knowledge spoon fed to them. Those are the same people that will balk at Python because it's not the exact same set of syntax and APIs that they memorized back in class.
Much of the "JavaScript on the Web" headaches people oft bitch about are entirely irrelevant with JavaScript on the Desktop. JavaScript can suck for Web programming because you have to deal with the least common denominator, which generally means IE6, which in turn means a horrendously out of date and horrendously buggy implementation of an ancient version of the JavaScript language.
Slightly back on topic, but related to my rant there, I will note that the idea that a JavaScript+HTML+CSS toolkit for GNOME will somehow magically attract more developers is totally bunk. Real applications need real developers. People who've never managed to learn anything besides JavaScript+HTML+CSS are not programmers and do not write applications. They write goofy animated web pages with some backend database integration. The people writing GMail for instance are not your common Web developer; compared to other Web developers, they are gods among rodents. Those aren't the kind of people who have been wanting to write GNOME apps but sat there and thought, "gee, I only know JavaScript, GNOME is too hard!" Those are the people who could probably code a GNOME app in pure assembler, but aren't writing GNOME apps because they just don't want to or just don't care.
Posted Jul 29, 2010 11:32 UTC (Thu)
by nye (guest, #51576)
[Link]
Every point in this post is spot on.
Posted Jul 30, 2010 0:41 UTC (Fri)
by ras (subscriber, #33059)
[Link] (3 responses)
Actually, Python is substantially faster than Ruby, Perl, PHP, and most JavaScript implementations for that matter. See: That said, there is one very fast JavaScript implementation out there - V8. Given the amount of time and effort being poured into JavaScript I expect it will be the fastest in the medium term. But not in the long term, primarily because Python and JavaScript are almost identical languages under the hood. The main difference is Python has a sane syntax, so you don't have to go looking for the good parts. Given V8 already runs 3 times faster than Python, I suspect the urge to apply the techniques used in V8 to Python will become irresistible. If you are doubting that Python and JavaScript are very similar have a look at Pyjamas, a Python to JavaScript Compiler. Turns out there is almost a 1 to 1 correspondence between Python constructs and JavaScript.
Posted Aug 2, 2010 15:34 UTC (Mon)
by nye (guest, #51576)
[Link] (2 responses)
In your opinion. I find Python to be distressingly inelegant at times, and has enough gotchas to be mildly dangerous. Reasonable people can differ here.
>If you are doubting that Python and JavaScript are very similar have a look at Pyjamas, a Python to JavaScript Compiler. Turns out there is almost a 1 to 1 correspondence between Python constructs and JavaScript.
The conclusion doesn't necessarily follow from the premise. Python can be compiled to JavaScript pretty successfully, but I think the reverse would be harder (Python doesn't even have closures; how can you take a dynamic language seriously without closures? :p) - although as we all know, it is possible *in principle* to compile any (Turing complete) language into any other.
This probably doesn't change your point about speed, and it would be interesting to see if the same techniques can be successfully applied to speed up Python, but the fact that JavaScript is improving in that area faster than Python would be a warning sign.
Posted Aug 2, 2010 22:49 UTC (Mon)
by ras (subscriber, #33059)
[Link] (1 responses)
Posted Aug 2, 2010 15:34 UTC (Mon) by nye (subscriber, #51576): Quoting from Wikipedia: Python has had support for lexical closures since version 2.2.
Posted Aug 3, 2010 10:37 UTC (Tue)
by nye (guest, #51576)
[Link]
Many more thanks for introducing me to the 'nonlocal' keyword - finally a good reason to use Python 3.
Posted Jul 29, 2010 6:31 UTC (Thu)
by magnus (subscriber, #34778)
[Link]
http://www.youtube.com/watch?v=hQVTIJBZook
Posted Jul 29, 2010 6:40 UTC (Thu)
by Cato (guest, #7643)
[Link]
As for 'slow' - there's a lot of work on JavaScript performance using JITs, which has resulted in some significant gains: http://www.codinghorror.com/blog/archives/000509.html. There is intense competition between iOS and Android in this area as well: http://arstechnica.com/gadgets/news/2010/07/android-22-de...
Posted Jul 29, 2010 4:16 UTC (Thu)
by thedevil (guest, #32913)
[Link] (2 responses)
Libre / Open Source was about users' freedom to fix or tweak the code they run to their satisfaction, or pay someone to do that for them. If the code runs on a server in a data center somewhere, how does the source code help? You can have the perfect patch but if a distant organization (profit or not) is responsible and not interested, the fix isn't going in. You can fork but you'll need your own server and your own rack in the data center.
Perhaps the idea is that today, or soon, anyone can get a hosting solution to run these great Web Apps of the Future. I think that is way too optimistic.
Posted Jul 29, 2010 8:09 UTC (Thu)
by mjthayer (guest, #39183)
[Link]
See also my comment further down. Today, at least I get most of my software from the repositories of a private company - Canonical Ltd. Nonetheless, the source code is available and I have fed at least one patch back, via upstream, to Canonical's repositories. (Not to mention the software I work on in my paid time, which is also sitting there.) As long as you can also run the software locally for testing purposes (and to be honest, testing a patched version of a system component that lies a bit deeper down the stack is a bit of a pain even with an Ubuntu desktop) you can improve it and contribute back your improvements. Then you just have to choose a provider that plays the game.
Posted Jul 29, 2010 19:11 UTC (Thu)
by alfille (subscriber, #1631)
[Link]
Say you want to write an application. Almost anything, say: computer algebra, screen writing, board game or sprinkler controller. You could write the interface in gtk or qt, or you could write it to the browser. Your backend runs it's own web server process, and you are instantly cross-platform, distributed and web appliance capable. Pretty compelling.
What is lost with the browser interface is the communication between subtle changes in your program state and the rest of the desktop, and communication between the browser windows. That's where gnome should concentrate.
Posted Jul 29, 2010 4:18 UTC (Thu)
by jmorris42 (guest, #2203)
[Link] (12 responses)
Is it because making everything a client server ajax app is going to be easier? Not bloody likely. Is it to make it network transparent? Uh, we have had that since day one with X. Cross platform? GNOME/Gtk is already more cross platform that most existing complex web apps.
Making every app aware of URLs, yea that is a good idea.... one that has been implemented to varying degrees of completeness for over a decade.
One thing I can promise. We already lost the netbook war because we were more bloated and slower than Windows XP. We make our desktop slower still and we can cede the tablet and smartphone game to Android and all just learn to write it's mutant flavor of Java.
Is the idea to abandon the local desktop for some cloud based thing? Again, Android is here now. Within another year I fully expect all major distros to support the Android Marketplace and Android apps running seamlessly on current desktops.
Posted Jul 29, 2010 5:31 UTC (Thu)
by robertm (subscriber, #20200)
[Link] (6 responses)
And unless something happens now, this is going to be a crushing blow to free software for those same end-users. It's gotten solidly entrenched on the server side of things and will probably remain there until a completely unpredictable game-changer comes along to shake things up, but for end users? Completely Tivoized appliances talking to proprietary applications.
And it's a pity, because the web is a mess. It's a horrifying hack atop horrifying hack atop wildly inefficient use of resources atop more horrifying hacks. It is the epitome of "worse is better" when it comes to winning marketshare. But as much as I wish it would go crawl in a corner and die a much-deserved death, it's here to stay.
Posted Jul 29, 2010 5:41 UTC (Thu)
by thedevil (guest, #32913)
[Link] (4 responses)
Posted Jul 29, 2010 6:30 UTC (Thu)
by Tara_Li (guest, #26706)
[Link] (3 responses)
Seriously, just this evening, I was having random inability to get to huge chunks of the Internet - at one point, even Google wouldn't come up for about 30 minutes. If I have nothing local - I have a paperweight on my desktop if my network goes down.
And honestly, the best way to keep a computer virus-free? Don't connect it to the Net!
Posted Jul 29, 2010 7:53 UTC (Thu)
by dlang (guest, #313)
[Link] (1 responses)
all these extra layers will work find if you are on a high-powered mains powered system connected to a high-speed, low-latency network connection. but change any of these criteria and the feasability of 'everything is the web' starts falling apart rapidly.
OLPC and the netbooks pointed out to people that there is a backside to moors law, namely the same capability gets cheaper over time. today a few people are willing to pay large sums of money for smartphones. when the same capability that sells today for $600 drops to $100 there will be a _LOT_ more of them around.
if you are on a high-latency network connection (say a satellite feed if you are in rural areas) every round trip to the server is very painful (approximately 1 second)
I already mentioned the battery life issue, radios are expensive to power (and as the power for the rest of the system drops over time, they become even more expensive as a relative cost).
Posted Jul 29, 2010 10:35 UTC (Thu)
by NAR (subscriber, #1313)
[Link]
Exactly. In my experience it takes about 2-5 seconds to open an e-mail in gmail while with the local pine client it's instant. WWW stands for World Wide Wait. In my experience the latency is simply unacceptable for most web based applications. Reliability is also worse: it's not enough that my computer works, the connection to my ISP must work, the international line of my ISP must work and also the server must not be down. A number of extra components that can broke down.
Posted Jul 29, 2010 7:56 UTC (Thu)
by mjthayer (guest, #39183)
[Link]
I am one of the most clueless people about web applications going around, but my understanding was that Javascript applications were stored on a remote server but cached and executed locally. Surely with a bit of tweaking (like making sure that the whole application gets cached and not just part of it) it could continue to work once your connection is gone. apt-get normally also initially relies on a network connection.
Posted Jul 29, 2010 6:35 UTC (Thu)
by jmorris42 (guest, #2203)
[Link]
Perhaps. Perhaps not. Google was dreaming of ChromeOS, where the browser would be the only local app. Android ate its lunch. Besides handsets there are lots of Android tablets in the pipeline now, probably netbooks soon. Besides the plain fact that a local app is going to outperform a cloud app there is the mercenary fees available when app stores sell downloadable apps that few would pay for access to the same app on a hosted basis. It might be irrational but it is real.
Posted Jul 29, 2010 10:01 UTC (Thu)
by droundy (subscriber, #4559)
[Link] (4 responses)
I don't know if you've tried running an X application over the network recently. I ran emacs yesterday across the country, and let me tell you, it was painful! We're talking minutes of startup time and perhaps 10 seconds to refresh the screen. I'd be surprised if that's what's going to beat AJAX.
On the other hand, I don't look forward to the bloat of making everything run under the browser, either...
Posted Jul 29, 2010 10:32 UTC (Thu)
by Frej (guest, #4165)
[Link]
Posted Jul 29, 2010 20:35 UTC (Thu)
by zlynx (guest, #2285)
[Link] (2 responses)
With Canvas and/or SVG, combined with Image objects and HTML fragments, an application could send graphics and text and receive keyboard and mouse events.
Lots of stuff wouldn't work, but I bet it could be very interesting.
Actually, isn't there already some version of ExtJS that can render for a Qt backend using signal/slot programming? Pretty sure I read about that somewhere...
Posted Aug 2, 2010 0:37 UTC (Mon)
by bronson (subscriber, #4806)
[Link]
http://kanaka.github.com/noVNC/
Posted Aug 3, 2010 0:51 UTC (Tue)
by roelofs (guest, #2599)
[Link]
There's a lovely version called XML11. It implements X11 (some level of it...) in JavaScript with XML messaging, and the demos, at least, were very cool. But that's not all it does: since it's JS-based, it can actually run code locally on the server (i.e., your browser), not just regular X display stuff. The canonical example is a calculator app; the "X11 version" just sends display commands to the browser and accepts events (button pushes, typed numbers, etc.) from it, with the calculations happening on the "client" end (in the X sense--what everybody else calls the server). But the "XML11 version" actually sends some or all of the application code to the browser/server to execute there. I vaguely recall that the X client ("regular server") code was Java-based.
At any rate, my experience with it started with a demo/presentation and ended with some reading of the docs, so I don't really know how complete it is. Looks like development ended about three and a half years ago--unfortunate but not surprising, given its relative lack of publicity. It might be worth resurrecting, though, if web-based X apps intrigue you.
Greg
Posted Jul 29, 2010 12:48 UTC (Thu)
by richmoore (guest, #53133)
[Link]
Posted Jul 29, 2010 13:17 UTC (Thu)
by mitchskin (guest, #32405)
[Link] (1 responses)
There may be a day when everything is in the browser, but that's so far off that it's too early to start focusing on that. What will that look like? How will it work? How does it relate to GNOME, as we currently understand it? People are barely beginning to experiment with answering those questions now; it will be a long time before it's a mass phenomenon. And if you start pushing hard on a particular approach this early, you run a greater risk of ending up in a blind alley.
It would be better to focus on where we are now, and get that right. The current GNOME codebase and its derivatives aren't done yet, and there's a lot of room for integrating with the web rather than moving wholesale to it.
And while we wait for the future to get a little clearer, we can spend some time exploring the possibilities. Maybe that exploration is what he was advocating in this talk? I went to the GUADEC site hoping to watch a stream of the talk but it looks like they only have live streams.
Posted Jul 29, 2010 14:40 UTC (Thu)
by jtc (guest, #6246)
[Link]
My guess is that in 10 years, "the browser" will be a very different thing than it is now. But, as your questions imply, it's not possible now to know what "the browser" will be at that time. But I think some educated guesses can be made. I suppose that those who guess right and act on their guesses will tend to be the ones who win out in the future. (And, hopefully, FOSS developers and proponents will be among those who guessed right.)
Posted Jul 30, 2010 14:41 UTC (Fri)
by tstover (guest, #56283)
[Link]
So if you want to jump in the fire with the superstition, anarchy, and decay that is the dark age of ajax, iphones, flash, and big brother in your firmware, then go have a good time - but get off my lawn. :) In ten years I guarantee people like me will still be making money with our beloved C, *nix, ipv4, and yes desktop computing. Not the least of which will be the ones developing everyone else's little sandboxed browsers they think are such a big deal.
(the more I pursue this character of senseless antagonism towards "web technologies", the funnier I think it is, and the more I love it!)
Posted Jul 30, 2010 16:40 UTC (Fri)
by robert_s (subscriber, #42402)
[Link] (1 responses)
Posted Aug 4, 2010 20:10 UTC (Wed)
by behdad (guest, #18708)
[Link]
Posted Aug 5, 2010 11:38 UTC (Thu)
by eduperez (guest, #11232)
[Link]
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
To think that Python is an answer because JavaScript is slow just proves you have abso-freaking-lutely no idea what you're talking about. Python is one of the slowest of all the popular interpreted languages out there by a good margin, and JavaScript implementations have easily taken the crown as being the fastest. Hell, even Ruby (once the absolute king of slow-and-stupid runtimes) is faster than Python these days.
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
Python doesn't even have closures
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
Whoosh!
Whoosh!
We turn everything into a web app... why?
Because, ultimately, the desktop computer is on the way out. Appliances smartphones, game consoles, DVRs are the future of the computer for end-users outside the workplace. The closest most people will come to a general-purpose personal system will likely be a tablet form-factor system running almost exclusively a browser of some kind. It's important to build applications which are web-aware because within half a decade the vast majority of non-business and a good-sized chunk of professional software is going to be running on a server that the user never sees or in Javascript on the browser.
Whoosh!
Whoosh!
Whoosh!
all these extra layers will work find if you are on a high-powered mains powered system connected to a high-speed, low-latency network connection. but change any of these criteria and the feasability of 'everything is the web' starts falling apart rapidly.
Whoosh!
Whoosh!
Whoosh!
Is it because making everything a client server ajax app is going to be easier? Not bloody likely. Is it to make it network transparent? Uh, we have had that since day one with X. Whoosh!
Whoosh!
Remote access protocols
Remote access protocols
http://sourceforge.net/projects/thinvnc/
http://sourceforge.net/projects/guacamole/
Thinking about running stuff over X over the the net, I wonder how difficult and strange it would be to implement a version of X that ran in a browser?
Remote access protocols
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
totally disagree
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web
GUADEC: Luis Villa points GNOME at the web