The Null Device
Posts matching tags 'open-source'
Google's Open Source blog has a video interview with Australian open-source developer Rusty Russell, who has contributed to gcc and the Linux kernel, in which he puts forward a possible explanation for why there are disproportionately many Australian open source developers; and explanation based upon that ubiquitous factor within Australian culture and society, the tyranny of distance. Russell contends that the fact that Australia is so far away from the centres of technological development means that young Australians with a talent for coding are more likely to express that by getting involved in projects on the internet, and, more often than not, open-source projects.
I wonder whether this means that areas with high concentrations of technological companies and startups (such as the Bay Area, London and Berlin) contribute less to open source than (internet-connected) places in the middle of nowhere.
A few years ago, a few geeks in the UK, displeased with the Ordnance Survey's hoarding of taxpayer-funded mapping data, decided to do something about it, and so OpenStreetMap was born. Based on the same principle as Wikipedia, it used the power of internet-based grass-roots organisation to allow people to make their own maps, walking roads with GPS units, uploading the traces and giving them names. Out-of-copyright vintage maps and donated satellite data, among other things, helped a bit.
As you can imagine, in the early days, it wasn't much to look at. There were networks of roads, though most of them weren't named, and a lot were missing. You could sort of make out where you were, if you knew the place well. The interface was also somewhat slow and clunky, compared to Google Maps (a variation on whose draggable-tiled-map theme it was).
Some are impressively comprehensive; for example, OSM's maps of Reykjavík (and, indeed, Iceland's second city, Akureyri) and Buenos Aires are as thorough as anything you'd expect from Google, were they to bother. Harare, whilst looking quite sparse, is more detailed than on Google's maps, and the Papua New Guinean capital of Port Moresby seems fairly comprehensively drawn. Even Pyongyang has a surprising amount of detail (though one probably can't blame OpenStreetMap for most of the streets being seemingly unnamed); I imagine that as soon as North Korea allows unrestricted tourism, long before the first McDonalds goes up, tourists will be dragging cached OpenStreetMap tiles on their jailbroken iPhones as they negotiate its broad Stalinist boulevards.
Being based on free data, OpenStreetMap has other advantages over its commercial cousins. Each map comes with an Export tab, which lets you grab the displayed area in a variety of formats, from rendered pixmaps to SVG or PostScript to the actual raw data. With it being under the Creative Commons, you are free to do as you like with the data (subject to a "share alike" proviso if you commercialise it). And with it having the agility of the Wiki age, OpenStreetMap is starting to steal a march on its competitors; for example, it was the first map to show Heathrow Terminal 5 correctly.
Of course, OpenStreetMap is by no means perfect. parts of the world are still uncovered (much of the Russian interior), or only covered with major roads (much of Africa). And their rendering algorithm doesn't seem to do Chinese or Japanese characters, rendering most of China's place names as rows of boxes. (If one is picky, one could request transliterations of foreign character sets; perhaps this could be done as user-selectable layers.) There is no route-finding capability (of the sort that Google Maps has). But all in all, OpenStreetMap is very impressive, and a spectacular success.
In case you haven't seen it: Radiohead's video for their new song House Of Cards. Shot using no cameras but only 3D volumetric scanners and computer rendering, the video consists of disintegrating 3D dot representations of suburban streetscapes, key parties and Thom Yorke singing. What's more, there's a video on the making of it, the data set is available for download (in CSV format), and Radiohead are soliciting user-made reinterpretations of this video here.
Remember BeOS, the super-nifty object-oriented operating system of the 1990s? Well, BeOS itself may be no more (apparently Palm bought it and then proceeded to not do much with it), but there is now an open-source BeOS-influenced OS named Haiku, which appears reasonably functional. There are images downloadable which you can play with in VMWare Player or similar. There's an article about it here.
Apparently Microsoft have promised to release the full specifications of their legacy binary Office document formats, making them available for direct downloading from their web site without the need to sign any agreements. Not only that, but to develop a reference application for translating them into a neutral format and release it under the BSD license. Cue a million Slashdot penguinheads trying to outdo themselves at saying "Hell has frozen over" in the wittiest way possible.
Sun has bought MySQL, maker of the popular open-source database system. Which looks like good news to all concerned, as Sun have a good reputation for supporting open source.
The source code of the classic urban-planning simulation game, SimCity, has now been released under the GPL. You can find it here. The code is based on the original, UNIX/X11/Tcl/Tk version of SimCity, with a few changes: (a) the game has been renamed to Micropolis (which was its original working title), as "SimCity" is an Electronic Arts trademark for their commercial urban-simulation games, (b) it has been ported to the OLPC XO-1 (the cute green laptop being given to children in developing countries), and (c) everything has been placed in a C++ class and bound to a Python interpreter, making the entire game eminently hackable and extensible in Python. Let a million hacks bloom.
Why is it, you may have asked yourself, that a technological civilisation that can put men on the moon, map the human genome and create the Nintendo Wii and the iPod can't make a standards-compliant web browser that doesn't leak memory like a sieve. Well, there's some good news on the horizon: the developers of Firefox have embarked on a memory leak eradication drive:
Aaron suggested having an "about:memory" page showing a breakdown of Firefox's memory use (bug 392351). When I pointed out the bug to Brendan Eich, he excitedly assigned the bug to himself.
Robert Sayre created a script to load random pages and see whether they cause leaks. The random URLs come from the Yahoo directory (biased toward older, top-level pages), del.icio.us (biased toward newer, geeky pages), and AltaVista (biased toward pornography).I see they have their use cases covered.
Steve England tested the top 500 web sites, finding two leaks. Later, he tested the top 20 Firefox extensions and found leaks in several of them.And there are some interesting user comments on the page.
Could I suggest a test of a 10 minute session of scrolling and zooming around in google maps hybrid mode as something guaranteed to to eat over 1GB of memory?I'd venture to say, from personal experience, that Yahoo! Maps (which appears to be a clone of Google Maps, and and is, to the best of my knowledge, only used for geotagging photos in Flickr) appears to chew up more memory than Google Maps. Which is rather funny, what with Yahoo! employing some of the brightest minds in AJAX development today (Douglas Crockford, for one).
Anyway, good luck to the Mozilla developers. Speaking as one in the habit of leaving lots of windows open in a session, I hope that this will lead to a browser that doesn't chew up all of the computer's resources if used for more than a few hours.
Last night, I went to the London Open Source Jam, a Linux-themed show-and-tell hosted at the Google offices near Victoria Station.
They had various speakers who came in, bringing in variously interesting Linux-based projects, talking (very) briefly about them and then letting people poke around with them. They had some people from the One Laptop Per Child project, who brought along two prototypes of the US$100 laptops which they are going to build and distribute to children in the developing world. The machines are very well designed (they're full of innovative design features, they're reportedly almost indestructable, and they look desirably good); they're based on Linux and Python, look nothing like the machines used elsewhere, and are designed to encourage tinkering and exploration. They even have a key (labelled with a graphic of a cog) which, when pressed, takes you to the source code for the currently running program so you can hack and modify it. (That didn't seem to be working on the laptop I looked at, though that prototype wasn't working perfectly.)
Someone also brought along a Trolltech Greenphone; that's a new mobile phone, created by Trolltech (the Norwegian company that makes the Qt toolkit (a rather elegant GUI library) and the Qtopia interface for portable devices), and based entirely on open-source(ish) software. It looked much like a regular phone, albeit with some developer features. I saw no evidence of it containing WiFi, though, so it may be lacking in that department.
Sony sent along a representative with a PS3 console, to show that it could boot Linux (and not only from special Sony-approved distributions, as the PS2 could, but from any distribution of your choice). Of course, the catch is that when the PS3 boots an untrusted Linux disc of your own providing, it runs it in a sandbox (under a Xen-style hypervisor), isolating it completely from the nifty graphics chip (so that game developers can't evade Sony's game-licensing fees by distributing full-featured games as bootable Linux live DVDs), and reducing it to a generic computer with a somewhat unusual CPU. They had a demo, which rendered a Mandelbrot set and let you zoom around on it; it was like a faster version of something I saw on a 386 PC in 1989. All in all, I found this demo underwhelming; when asked why anyone would want to run Linux in a sandbox on a PS3 when one could build a PC for less and get access to better graphics capabilities, the Sony rep didn't have a good answer; it was intended, she said, for people who already have a PS3 and, for some reason, are seized with the desire to run Linux on it. Though I suspect that, the fact that doing so doesn't involve breaking locks or using the hardware for anything more than it was designed for (the fact that the PS3 can be a mediocre Linux box is about as exciting as the fact that an iPod can be a mediocre PDA, or a mobile phone can be a mediocre MP3 player), will largely leach any such endeavour of any hack value it may have had.
For those looking for the ideal present for the hacker in their lives, Make has an open-source gift guide. This includes all sorts of nifty (in a rather geeky way) things, from a DIY TB-303 clone kit to a software radio transceiver that can handle all kinds of signals to persistence of vision displays for bicycle wheels, and a DIY game console (on which you can play and write 1982-style video games), a Linux-based pocket game console, and not one but two open-source mobile phone platforms, not to mention numerous controller and interface boards to build stuff out of.
Sun have announced that Java will be available under the GNU General Public Licence. Presumably because Flash/Python/PHP/.NET were eating their lunch. It remains to be seen whether this will prolong Java's life, or results in some of the nicer bits (such as the class libraries) being salvaged and bolted onto more vibrant platforms.
Hot on the heels of the TrollTech Greenphone, a company named FIC is releasing an open-source, Linux-based mobile phone. The OpenMoko won't have WiFi (as the GreenPhone will), but it will have a GPS receiver built in, as well as Bluetooth, USB, a MicroSD slot and a multi-touch touchscreen capable of understanding two-fingered dragging gestures. More importantly, it's not so much a mobile phone (which is to say, a locked-down proprietary appliance) as an openly hackable general-purpose computer with a GSM module and GPS receiver attached. (The radio components are, for regulatory reasons, closed modules; everything else, though, is fair game.) And apparently it will have a Debian-style package management system built in which can download new (or updated) software components on demand. And, also unlike the Greenphone, it will be released to the mass market (they say in January, which could mean sometime in the first half of 2007).
It looks like the Rockbox open-source MP3 player firmware is starting to get noticed. After a few years of hobbyist hackers independently shoehorning it into running on various players, often in spite of the manufacturers, one player manufacturer, SanDisk, is reported to be negotiating officially porting it to their hardware and supporting it. Which could be very good news, especially if they contribute their port back to the project, ensuring that it can keep up to date with developments.
Unfortunately, being the product of a flash memory manufacturer, the SanDisk players are all flash-based, and peak at 6Gb (though do have SD-card slots, so you can add an extra gigabyte or so if needed). Now if someone made a 40-60Gb hard-disk-based player that ran Rockbox by design, I'd be interested.
Open-source messaging client Gaim is about to get a raft of new features, bringing it into the 21st century. Main Gaim developer Sean Egan has been hired by Google, and part of his job is integrating Google Talk (and thus voice functionality) into Gaim; as part of this, the gaim-vv branch will be merged back into Gaim proper, also giving video/webcam support. Other changes in Gaim 2.0, expected in two months' time or so, will include less broken file transfers and a bunch of new IM protocols, one of which will be "Apple's Bonjour" (by which they presumably mean serverless chat with people on a nearby network).
Instant messaging on Linux is about to get somewhat less sucky (though, with any luck, they'll leave out the full-screen Flash spamming capability MSN is getting). And, with any luck, the changes will end up in Gaim-based clients for other platforms, such as Adium for OSX.
Everybody Loves Eric Raymond, a web comic for penguinheads, involving Richard M. Stallman, Linus Torvalds and Eric S. Raymond living in a shared house.
And then there's Buzz Aldrin's Conspiracy Smackdown:
The EU Parliament has thrown out a proposed software patent directive, by 648 to 14 (w00t!). The European Commission has said that it would not draw up or submit any new versions of the proposal. Which means that it is stone cold dead, for now at least; though as Cory Doctorow points out, there is too much monopoly rent waiting to be extracted for the pro-patent lobby to not try again.
(via bOING bOING)
I have been thinking about the homebrew-console-games-vs.-manufacturer-DRM issue recently.
New consoles with new capabilities come out, often containing powerful CPUs and graphics chips, and hackers and hobbyists want to have a go at writing code for them and getting them to do things other than consume titles. The manufacturers, of course, design the units so as to prevent unauthorised code running on them, primarily to protect their business model. The video-game console business model typically involves selling the consoles cheaply (often at a loss) and collecting a cut of the price of each game sold. Of course, for this to work, console makers have to strictly control what code will run on their machines, and ensure that they get a cut of every item released for them.
It's a stiflingly regressive reality, though it appears to be stable and is unlikely to go away any time soon. The alternative model (open game machines, sold at above-cost price, with anyone able to develop code for them) has been tried and failed; witness the Tapwave Zodiac PalmOS-based game machine, for example. Customers are more likely to buy cheap consoles and more expensive games for them later, in instalments, than to buy a more expensive console with cheaper software. Of course, this makes game consoles somewhat stagnant platforms (compared to, say, PCs or handhelds), though the game market seems to be able to cope with this well enough for it to be the best current business model for that kind of business.
(This ignores mobile phone J2ME games, which anyone can write and run on any compliant mobile phone without the manufacturer's blessing. Mobile phones are heavily subsidised as well, though they are subsidised by phone companies who make the money back in network usage; besides which, J2ME is a fairly weak gaming platform (for one, the low-power CPUs used in mobile phones often mean sluggish response times for navigating the internal menus, let alone games). Perhaps this will change in future.)
Nonetheless, that does not change the fact that hardware such as the PSP and Nintendo DS is tantalisingly attractive to tinkerers. When it was discovered recently that certain early Japanese PSPs could be made to execute code off a Memory Stick, a hacker community cropped up, with games, demos, utilities and ports of old console emulators popping up like mushrooms after a rain; the more recent firmware has closed off this hole, and anyone running a recent game on an old PSP will find themselves upgraded against their will.
What if, instead of locking out the hacker culture, game companies worked with it, whilst still preserving their business model? Imagine, for example, a device sold by the console manufacturer which costs about the difference between the retail and cost price of a game machine and enables it to run homebrew code. It could be a disc, a hardware dongle, or even a special cable. Unlike homebrew hacks (such as the Nintendo DS passthrough cartridge), it requires no soldering and no fabrication of circuit boards, allowing those who don't have a fetish for that sort of thing to get involved. Perhaps it comes with development tools and documentation (the GNU toolchain would be a start), or even membership of a community web site, where users can share their code. From time to time, publishers could release compilations of the best such titles, perhaps in a magazine format, doing the necessary licensing to make the releases run on standard machines.
Sony once tried something like this with their PlayStation 1; they called it "Net Yaroze", and apparently it wasn't a stellar success. I wonder whether it could be done better.
Of course, if the console makers don't throw a bone to hobbyists, makers of third-party extensions (of various levels of legality) just might; and these would be less concerned with protecting the makers' profit margins.
In Australia, unemployed hackers can now fulfil their work-for-the-dole obligations by contributing to open-source projects.
The Rockbox iRiver porting effort is making some progress. They now have it booting on an iRiver H140 and presenting the file browser that will be familiar to Archos Rockbox users. Not only that, but someone has coded a GameBoy emulator plug-in, which the iRiver is powerful enough to run.
Now all they need to do is make it play sound; unlike the Archos, the iRiver has no hardware MP3 decoder, so they need to graft an entire codec API onto the system. With such a big step, of course, come plenty of possibilities; already there is talk of having the unit play tracker modules and Commodore 64 SID chiptunes, not to mention lots of lossy and lossless audio formats.
Hopefully they'll get it working on the H3xx series (that's the newer one, with a colour screen and two USB ports) too. It'll be less annoying than the stock iRiver firmware.
A man in London was jailed for using the Lynx text-mode browser to make a donation to the tsunami relief fund. Anti-fraud monitoring interpreted the unusual browser signature (i.e., not Mozilla or IE) as the sign of a hack-attempt and the police went in in a SWAT-style raid, smashing his door down and arresting him. Which goes to show that it does not pay to use unusual software.
IBM have turned over 500 software patents to the open-source community. IBM will continue to hold the patents, though have pledged not to assert them against software distributed under an OSI-approved open-source licence. (It's legally binding, too, so there's no possibility of a change of guard at IBM reneging on it.) They have, however, reserved the right to assert them against anyone suing open-source projects for patent infringement; i.e., those who don't get with the programme may find themselves out on a limb.
"The 500 patents include U.S. Patent number 5,185,861, registered in 1993, which covers technology that helps microprocessors use their memory caches efficiently; and U.S. patent number 5,617,568, registered in 1997, for allowing non-Windows based systems to act as file servers for Windows-based clients, according to IBM Asia Pacific spokeswoman June Namioka. Other examples include patents related to handwriting recognition, she said."
"There's little argument that over the past dozen years, the world has come to view things differently: free software is one aspect of this; globalization of trade is another; both have been profoundly influenced by access to the Internet and the Web, and the easy access to information they provide. Knowledge is, indeed, power. As the models change, people who are stuck in the older mode, like Gates . . . look increasingly like Pope Urban VIII and rms looks more like Galileo: despite 'common knowledge' the world was moving. IBM's freeing-up of patents is another step toward proliferating knowledge."Which is a good start; perhaps a neo-Galambosian world where all concepts are privatised and monetised isn't inevitable after all. Mind you, we're not out of the woods yet. The existence of software patents (in the US and Australia, at least; the EU has so far managed to escape this fate) still creates a minefield which threatens to take down anyone without an extensive patent portfolio of their own, cross-licensing agreements and a hefty legal department, and threatens to establish an oligopoly on software development and invention. Though, hopefully, the establishment of a "patent commons" of valuable patents, free to use for open-source projects but defensively assertable against those threatening such development, may act as a deterrent.
It looks like the people behind the Rockbox* MP3 player firmware/OS (which runs on Archos Jukebox series hardware) have started preliminary work on porting it to iRiver devices; they're starting on the (discontinued?) iHP-1x0 series, but hopefully that'll lead to an H-3x0 version as well (assuming that they're based on a similar architecture).
* not to be confused with Rocbox, a MP3 player brought out by rap record/fashion label Roc-A-Fella. (Hang on, aren't they part of Def Jam/Universal? If so, I wonder if their corporate parent knows that they're putting out a MP3 player.)
Not all that long after voting to adopt software patents, the EU are moving to legally require currency detection code in all image-processing software. This looks likely to either (a) be utterly ineffective, or (b) be mostly ineffective whilst effectively outlawing open-source graphics software. The precedent it sets is not a good one either; how long until paracopyright enforcement is mandated to be built into anything processing audio or video data, or indeed any copyrightable data?
Meanwhile, British Telecom have taken steps to block access to child pornography websites. A laudable sentiment, though one worries that the site-by-site censorship infrastructure required to implement this could easily be extended to blocking other things (overseas news sites publishing things violating the Official Secrets Act, for example, or MP3 download sites that piss off the local recording industry). One brave step towards the Singaporisation of the internet.
Meanwhile, the RIAA's latest campaign to defend the foundations of capitalism from the enemy within will involve putting fingerprint readers into music players to ensure that nobody who didn't pay for music gets to listen to it. Welcome to the Digital Millennium; make sure you've paid your way.
And the latest entry in the annals of hobby operating systems: MenuetOS, an OS whose main feature is that it's written entirely in assembly language. The product of an intrepid cabal of European hackers/masochists, it has a GUI that looks sort of like Windows XP would look if it didn't have proper widgets and was limited to one pixel font, and all the apps anyone in the target audience would need (i.e., it has an IRC client, a NNTP newsreader, MP3 playing/streaming capabilities, a chess client and a Doom-style 3D engine demo), and fits onto a single floppy.
Video Toaster, the classic Amiga video-editing system, has been released to the open-source community. Mind you, since what made it so useful is dependent on the Amiga's video chipset, it's more a historical curio than anything else.
Hewlett-Packard declare war on copyright violation, commit to integrated policeware in all HP products. Perhaps they're hoping that this gets them the sort of legal protection from cheap, usable imports that Microsoft has; since MS have embraced "trusted computing", the Microsoft OS monopoly has gone from being a problem to being an essential part of global economic stability (in fact, it's only slightly far-fetched to believe that, if it crashes down and the world goes open-source, the powers that be will declare capitalism to be irrecoverably wounded and bring out the cobalt-tipped fireworks). Could HP be angling for legal restrictions on non-compliant computers to stem the flow of customers to cheaper no-brand Asian imports?
Meanwhile, new versions of Photoshop have filters that detect images of US currency and refuse to load them; this CPU-intensive operation is performed every time image information is imported from outside the application. However, honest, patriotic citizens won't mind their CPU cycles being used in this way, as we all must put in to fight terrorism; don't you agree, Citizen?
Apparently Paint Shop Pro does this too, or so I heard. Mind you, The GIMP doesn't, and even if it did, you could change the source code to bypass this check and compile it yourself. Though, if Adobe and the Paint Shop Pro people put this "feature" in because of government pressure, it's not unlikely that there will be attempts to criminalise the distribution of source code that could be compiled to make a non-compliant image processing application. (There are precedents, in the FCC Broadcast Flag amendment, which effectively outlaws entire classes of software-radio applications that could be used to access copyrighted HDTV content.)
"Upgrading" is as simple as changing a version string. We already have it updated in Gaim 0.69. This was a no-brainer easy-to-fix thing, as was MSN.
If any Slashdot staff are watching, please, please refrain from posting articles related to IM unless you consult someone who knows what's going on. Too many trollish comments occur, and we get too many questions in Gaim support, all pointing at Slashdot as their source for the inaccurate information as to what's happening in IM.
Though wasn't it implied that from October, MSN's servers will require clients to produce a license certificate of some sort, which identify the client as a Microsoft-approved one, prior to connecting?
(If you don't want your IM network to be at the mercy of a profit-oriented corporation whose management may at any time decide to maximise profits by asserting control over your client, there's always Jabber, an open, decentralised, XML-based messaging system. Though nobody seems to be using that; I know of only one person on Jabber. Maybe if someone came up with some cute smiley themes for it...)
An interview with Richard M. Stallman, the head of the Free Software Foundation. In it he states his opposition to the Intel/Microsoft "Trusted Computing" system (or "Treacherous Computing", as he calls it), calls for web browsers to automatically send complaints to webmasters about Flash-based web pages (which, when you think about it, is not such a bad idea), and reveals that Debian has fallen out of favour because they tolerate the existence of a ghetto of non-free software. As well as the usual broadside at the Open Source movement (who, in rms's view, have more in common ideologically with Microsoft than with the Free Software movement). (via Slashdot)
After 9/11, governments quickly pulled formerly public information and restricted areas of scientific publication to keep them out of the hands of terrorists. Now DARPA, the US Department of Defense's research funding body has cancelled funding for OpenBSD security research because open-source software could help terrorist nation-states. Is this an isolated incident, or the start of a governmental purge of open standards and open-source software, and the start of a "national security"-driven shift towards proprietary standards kept on a strict need-to-know basis? After all, if Cisco, Microsoft and TRW hold the keys, the reasoning goes, Saddam Bin Laden can't use the technology to kill us. And replacing publically documented standards and open-source software with secret black-box technologies has numerous other advantages, from surveillance hooks to catch more terrorists, paedophiles, tax cheats and miscellaneous troublemakers to tremendous "peace dividends" such as end-to-end copyright enforcement and whistleblower-proof rights management for documents; not to mention handsome dividends for the shareholders of the keepers of the keys.
I just downloaded and compiled Sodipodi, after seeing a link to it from a Slashdot forum on SVG. It looks like a fairly promising vector graphic editor for Linux; and about time too. (KDE's Kontour is slow and bug-ridden, the new KDE Karbon requires KDE 3 which doesn't play well with Debian yet, so I haven't had a chance to try it.) It's still a bit rough (it prints lots of diagnostic messages on stdout, and lacks a few things, such as a "select all" option), but it's looking quite promising.
Software for reading Canon RAW image files under UNIX, and converting them to PNM files.
Guido van Rossum wins a FSF Award, for developing Python. And well deserved, too.
Good stuff: The BBC are experimenting with streaming programmes in the Ogg Vorbis format, an open-source, completely free format available on virtually all platforms (and not encumbered by licensing agreements, patents or other proprietarian evils); what's more, the Ogg developers have taken a stand against the copy-prevention trend. The BBC appear to be the first major broadcaster to use this format. Use it (and if you do, write in to let them know you do) before Bill Gates has a word with Tony Blair and it gets squashed by a directive from on high. (As happened with non-Microsoft systems in most parts of Britain's government.)
Some academic ratbag types in France have released an open-source DVD player, which apparently does CSS decryption. This is currently legal in France, though with the EU Directive on Copyright, it won't be for long. (via Slashdot)
Some enthusiastic teenagers have decided to create an open-source Windows-compatible OS from the ground up. So far they have a web page and a request for startup/shutdown graphics. No word on actual code yet, but this looks set to follow kick-ass vaporware Freedows into the realm of penguinhead legend.
Good rant about the surfeit of open-source projects: (Freshmeat)
Is there a cool KDE application? Rewrite it from scratch with the GTK+ toolkit. Don't contribute enhancements to the existing project. Don't even evaluate their codebase for potentially reusable code; that goes against the spirit of COMPETITION. You need to write something... cooler. Is there a Java version yet? Rewrite it. What about a console version? Rewrite it. A applet? Rewrite. An EPPLET? REWRITE. A pure assembly version? RE. WRITE. ... Remember that rewriting is always a more creative process than porting, and there's nothing more important in mental health than a creative outlet. Doing a search on the appindex for "icq" will demonstrate how popular this method of project development/psychotherapy can be.