International OSS Desktop Conference aKademy 2004

Torsten Rahn writes "The KDE Project is pleased to announce the successful completion of the KDE Community World Summit ("aKademy 2004") in Ludwigsburg (Germany) taking place from August 20th to 29th. With more than 230 KDE core developers, usability and accessibility experts, translators, editors and artists participating, the event is expected to have a huge and lasting impact on the next major releases of the leading Linux and Unix desktop environment. In addition, 270 visitors from the KDE user base and from other Free Software projects brought the total number of attendees to 500. The international participants, coming from 5 continents, took part in 65 talks, 10 full-day tutorials and numerous BoF-meetings over the course of 10 days. Thanks to this huge turnout and the numerous activities, the event evolved into the largest conference ever held that focused on a single open source desktop environment."

Comments

still 10x slower than BeOS

[kad77]

Forget flamewars... How about some efficiency standards?!?

Why is it that this candy-coated windowmanager runs like a *DOG* when it's just moving windows and drawing text on a 512mb 550MHz PIII system, and BeOS 4.0 (pre)release could run multiple video streams effortlessly without lag (may as well mention almost instant boot) on a 166Mhz PPC 604 with 128 MB RAM? 5 years ago.

Maybe getting paid for your work and quality go hand in hand in some products?

Re:still 10x slower than BeOS

[be-fan]

Well, the first thing to remember is that good technology doesn't ensure success. Indeed, the general rule is that the first product to be "good enough" is the winner, not the "best" product.

Now, there are a couple of reasons why BeOS was so great:

1) It was pervasively multithreaded. Each window had it's own rendering thread that ran independently of the application logic. This allowed apps to be very responsive, even under heavy load. It's sad that even on my 2.0GHz P4, Mozilla still blocks the UI for several seconds when certain pages take too long to load or render.

2) It had a phenomenal scheduler. It wasn't comparable, in many ways, to Linux's O(1) scheduler (it wasn't very scalable), but it was wonderfully optimized for interactive use. It's interactivity estimation was lightyears ahead of what's in the O(1) scheduler now.

3) It's multimedia subsystem was very good at moving data around the system efficiently.

4) It's toolkit was well-coded with respect to smart redraw and resize behavior.

Interestingly, the OS wasn't all that structurally different from Linux. It had a fairly POSIX-complient modular kernel. It run it's video and audio subsystems as a seperate process (like X, aRts, and Jack). It was just very well-implemented with an eye towards a fast and elegant UI.

Of course, the OS had it's darksides. The toolkit wasn't font-sensitive and layout-based, the VM was antiquated, I/O and network performance was only decent, etc. However, for the average desktop, this really wasn't a big deal, and not something that couldn't have been fixed had BeOS survived to today.

Re:still 10x slower than BeOS

[kad77]

Thinking back, I was asking that rhetorically. However, I am really glad that someone else that used/uses this wonderfully designed and very well thought through OS decided to recite some specifics that made it so.
That's why we come back to Slashdot!

I'd also like to point out, that in the interest of fostering development, once I signed up with the company, without cost to me they sent pressed OS releases very often, updated MetroWorks compiler and toolkit releases, and even some clothing!

Would Be, Inc. be influencing linux to become more modular, multi-threaded, and bulletproof (like QNX) if they were still around, given many popular linux apps would be running on it today?

Re:still 10x slower than BeOS

[IamTheRealMike]

It's worth pointing out that the pervasive multi-threading of BeOS was arguably as much a weakness as a strength.

Multi-threaded coding is hard and don't any coding jocks tell you otherwise. It is significantly harder to do 100% correctly (and nothing less than 100% correct will do) than single-threaded coding.

Indeed, according to one of the ex-Be engineers one of the things that hurt BeOS was that writing software for it was quite tricky, it basically meant writing robust thread-safe code even for a simple text editor. There's a good discussion of it here.

Anyway it's sort of academic. One of the main uses of multithreading in BeOS according to be-fan was to do window rendering in a separate thread. Linux will get something very similar within a few months when X compositing lands. OK so it won't be a thread inside the same app, it'll be a separate process which is rendering the entire screen at once but the effect is the same - no matter how busy the app is, you won't be able to "rub out" the contents of the window.

Re:still 10x slower than BeOS

[MrHanky]

But one reason why BeOS was so snappy was because apps also took advantage of its pervasive multi-threading and the scheduler. This was, AFAIK, because the API sort of forced you to write multi-threaded. So Be had a lot of small, snappy, but crappy, applications that you could use to show off its responsiveness. But most of the serious apps had to be ported, and then the advantages just disappeared.

Have you tried running Mozilla on BeOS? I mean, not a rotten old port of M18, but a recent build of Firefox. Yes, it's still being developed, and it's been getting a lot better. But its responsiveness is a lot worse than under Linux (not to mention the font rendering). Even when comparing a 266 MHz, 64 MB Powerbook running Linux to a 1.4 GHz, 256 MB Athlon XP running an RC of Zeta.

But no, the PB can't run multiple videostreams. It has problems running just one. But video codecs have become much more demanding since the sub 200 MHz days. And so has computer use in general. Net+ was a small and simple browser, but it's practically useless these days, like all small and simple browsers are. You need features to do stuff, and stuff takes up space. That's why a small and simple OS is hopelessly outdated, and big and bloated environments like KDE (500 MB for a basic install?) and OS X (~2.5 GB for a basic install) are the future.

neat

[NanoGator]

" the event is expected to have a huge and lasting impact on the next major releases of the leading Linux and Unix desktop environment."

I personally hope they are all having a good hard look at Apple's stuff. The main reason I'm not running Linux is that there's a lot of choice out there, and it shows. I hate running to Google every time I want to do something simple. Despite my many years of using Windows, I had no trouble using a Mac when the need arose.

Anyway, sorry if that sounded like a rant. I'm just hoping some of the work that comes out of this gathering deals with the end-user experience. I'd love to get away from Windows.

Re:flamewars? doncha have something better to do?

[3seas]

to those comments refering to dcop and dbus...and even corba

Yes guys, I'm aware of these, however neither are standard in the general scope of linux or FSF software but are instead desktop specific or to complexicated, and this only helps the divide of flamewars.

for whats it worth I even have a bounty set up for anyone who wants to create a bridge between linux and AROS (Amiga Research OS -- a FOSS project on sourceforge -- that can run hosted on linux. bounty thru Team AROS) via such existing projects as dbus ... or dcop... or hell, a bridge that will allow communication to/from AROS to linux IPC whatever used.

but ease of use and ease of adding to existing open source applications (this IPC port) and documentation of what functionality is accessible thru the port in such applications, is needed.

perhaps there is something to be learned from how Amiga did it, and it became standard, easy to impliment and use and generally the apps include documentation of accessible functionality.

dcop only deals with a small percentage of available packages... and dbus currently even less, but not even a handful of apps.

Standard doesn't mean having numerous and obscure way of doing it (IPC at the user level), as that is quite the opposite of standard.

So.... there are better things to do than flamewar over destops... as such is only a non-productive distraction.

its important, I cannot stress that enough, given what MS is up to in teh direction of "software factories" methodology --- their book has missed two publishing dates so far but they are doing what they do instead.... collect feedback on the scope of this direction via shorter articles, websites dedicated to software factories, etc...

google on - software factories MS book - and if you really understand what they are up to, then you too will realize the importance of getting an easy to use and impliment standard IPC in use. ... shrug...

Maybe that will be dcop or dbus.... A plan, good, fair or poor, is better than no plan at all. Being destop specific is not a plan for the bigger picture or scope of FOSS packages.