Well what about this, from the April Ximian Gnome 1.4 announcement?
Supported Platforms
-------------------
(snip)
Support for Debian "Woody," SuSE 7.1, Mandrake 8.0 and Solaris is forthcoming in the next few weeks. Sit tight; we haven't forgotten you.
You can read the full announcement yourself. In the meantime, we've had to do without or change desktop:(
(Sidenote: the lameness filter rejects... on a line by itself !!! That's just fantastic)
Feel free to debunk it. Explain why it's better for developers (and the user experience) to have to work out how to optimise to a new pipeline every couple of years, rather than to squeeze every last drop of speed out of one design before moving onto the next on (e.g.) a five year cycle.
The whole point of modern processor architectures is for the work of pipeline optimisation be done by either the processor or the compiler. The goal is less work for the application developer. This coincides with the trend towards higher level programming languages - I can't think of any large major applications architected such that the code needs to be hand optimized at instruction level. Sure, after profiling some parts may be tuned, but the size of applications today are just too difficult to design in overview with that scope. A huge amount of the transistor budget these days is taken by microcode that performs these optimizations on the fly, but with architectures such as Itanium you're seeing a move towards compile time ordering. But now I'm getting sidetracked...
Ever looked at the specification of a Playstation [e-scapegames.co.uk] and wondered how on earth developers got it to do what they had it doing by the end of its lifecycle?
This is a time-honoured trend in closed-hardware systems. Look at the last generation of games on the SNES. Look at the scene demos being put out for the Amiga in the mid 90's, essentially a decade old hardware platform at that point.
Ever wondered why early Playstation 2 games bit the weenie?
All first generation titles do not harness the full capacity of a machine; I think this is your point. But on open systems developers don't have to optimize anything, thanks to Moore's Law. Maybe it doesn't strictly live up to the "small-is-beautiful" aesthetic, but software development is about optimizing results. Time will be spent where the greatest payoff is, and since performance boosts are a natural consequence of progress more resources are devoted to development.
Yeah, Luxo Jr was a classic. It's amazing to see how far hardware has come; Luxo was rendered in realtime by a GeForce 3 card in a Macworld demo. I'd link up the demo but according to this interview it is not being released:(
I really don't want to sound like a zealot or anything like that, but I would like to know if the integrated peripherals (sound, video, NIC) will be supported under Linux. It seems like a no-brainer for the video, given NVIDIA's commendable (though closed) Linux drivers. I remember jumping out and buying a TNT2 when they announced Linux support for the card, then waiting a year for the 4.X drivers to finally arrive. Now that I'm finally going to upgrade my computer, the nForce features are attractive, but I would like to know they are supported at the time of purchase instead of having to while away months wishing I'd bought products with existing support.
Really, this isn't a whinge about NVIDIA. Any company that supports Linux to the parity of their Windows releases deserves respect (are you listening ATI? Cirrus Logic? Guillemot? Creative?) After a few initially iffy driver releases, the very latest versions have been pretty much perfect for me. I just want to know if the features of this chipset will be supported off the bat this time. A kernel module for the NIC is probably not a major issue, but a closed sound-card driver will be hard to reconcile with the ALSA project, set to be the new Linux sound standard in 2.5.x. But any support is better than none, as they have proven in the past.
I haven't seen it pointed out yet, but GDK/GTK 1.3 have had AA enabled for a while now, so this is very much an interim thing while we wait for the big gnome 2.0 release.
I've tried it out a bit and generally liked it. There are some problems with font sizes in certain applications, where the font is now larger than the widget, but then again this may be due to the changed font preferences required. It takes a bit of fudging the configurations in Debian, and make sure you have a symlink/etc/X11/XF86Config to your/etc/X11/XF86Config-4 if you're running XF4.0, or the config script dies.
The GTK anti-aliasing is still being handled by the FreeType engine, which is IMHO perceptively as good as it gets. But you're begging for the screenshots aren't you? Here are some tiny morsels for you:)
Konqueror (and the rest of QT) has xft enabled for a while now - shot 1
And here is the gdkxft working on the gnapster menu bar - similar results. shot 2
That's twisting thet truth significantly. The major reasons cited in the Wired article for the failure of Linux adoption were:
Problems with hardware compatibility with theexisting computers in use. This would not apply if compatible hardware had been initially targeted upon purchase of the system, as it had been done with Windows.
A lack of local Linux expertise among teachers. Just because the teachers are not experienced with using Linux as opposed to Windows does not draw the conclusion that Linux is more difficult to use - just that the knowledge base is not there yet.
Political movements from the Government. Of course proprietary vendors weren't going to just sit back and lose out on a contract this large.
Furthermore, it appears the ScholarNet initiative is not over yet. The current progress has not achieved the penetration desired, but future iterations hold more promise, as hardware compatibility improves and the "seeded" Linux knowledge from the successful installations trickles down.
Anyone interested in the attractions of Linux implementations in developing countries might want to have a look at a paper I wrote for a final year Engineering unit: postscript version. It has some mistakes in it I haven't corrected yet, but I'm open to revising it:)
...it was ported to the PS/2 with the 1.2.1 release a few weeks ago. The port was done by some Japanese coders. The news here is that Sony has donated a dev platform for further development...so we can all hope Mesa is ported sometime:)
This is 100% correct. Not only that, after they changed it, we were told we couldn't change it to something less obvious for a few days while they "worked out a few kinks".
Telstra Bigpond has been unequivocally the WORST internet service I've ever used, of 4 ISP's tried.
As a BigPond ADSL user I have much to be thankful for. Thanks for the two-weeks downtime last month, changing the user agreement on download restrictions after the contract was signed, and forcing me to call every 4 days to reset my account when an authentication error on your end hangs the connection.
But most of all, thanks for leaking the account passwords through poor security and having the foresight to keep the server down right now so I can't change mine.
Re:Some corrections to the "article"
on
DeMuDi Linux
·
· Score: 1
I believe the low-latency patches give latencies of the order 1-2ms. RTlinux is overkill for this type of problem, and requires specialised hardware support in the form of new drivers for the soundcards. AFAIK, no commonly used cards have RTlinux driver support...and there are no audio apps that use the RTlinux API either.
More clearly, if you actually read the GPL that comes with Linux it has a clause explicitly permitting linking of non-GPL code with the system, which is considered as "normal use".
Corporate licenses of XP do not have the "fingerprint" restriction, which is why corporate copies of the OS and presumably Office will be the most coveted and pirated of all. Regular joes will end up footing the bill.
I can't wait till my ISP contacts me saying "The RIAA wants us to shut you down because you run a Freenet server" my response will be "Um - ok, whats their beef? Oh copyrighted music? Prove that data lives on my server! What? Its encrypted? Wow. Do I have the keys? Maybe - maybe not. I have no idea. Oh yeah - thats right - you aren't liable for criminal traffic going through your network either if you can't identify it? Well neither can I." Click.
I'd be more like "Encrypted copyright materials on my server? That's really perturbing...I tell you what, I'll pay you $10000 to figure out what the material is...". Then it's DMCA - pow!
Understood. I wanted to know if any research has been done to determine the discovery time and whether this would make the scheme feasible...although I'm pretty sure no, for the same reason cryptography is secure these days.
A while ago on slashdot.org an article was posted about a prime number that had the "magical" property of being able to be g'unzip'd to the DeCSS source. The number was represented as a (relatively short) arithmetic expression.
My question is, has anyone put any efforts towards seeing if other, larger pieces of data could be represented in this way? Of course the chances of finding these type of numbers are very slim, but they *are* infinitely large spaces, and the compression would be ultra-compact.
Maybe something for a quantum computer to chew on anyway...
Hey, Digi-Paint wasn't written in Assembler. [Digi] was the original HAM mode paint program, and was pretty slow at the time (but still useable). You're thinking of Brilliance, the wicked fast paint package that took, I believe, 5 man years of coding.
Still, DPaint really did cane, although it was slower towards the end. I'm amazed there isn't a paint program like it for Linux.
Slashdot Mirror
Well what about this, from the April Ximian Gnome 1.4 announcement?
Supported Platforms-------------------
(snip)
Support for Debian "Woody," SuSE 7.1, Mandrake 8.0 and Solaris is forthcoming in the next few weeks. Sit tight; we haven't forgotten you.
You can read the full announcement yourself. In the meantime, we've had to do without or change desktop :(
(Sidenote: the lameness filter rejects ... on a line by itself !!! That's just fantastic)
Feel free to debunk it. Explain why it's better for developers (and the user experience) to have to work out how to optimise to a new pipeline every couple of years, rather than to squeeze every last drop of speed out of one design before moving onto the next on (e.g.) a five year cycle.
The whole point of modern processor architectures is for the work of pipeline optimisation be done by either the processor or the compiler. The goal is less work for the application developer. This coincides with the trend towards higher level programming languages - I can't think of any large major applications architected such that the code needs to be hand optimized at instruction level. Sure, after profiling some parts may be tuned, but the size of applications today are just too difficult to design in overview with that scope. A huge amount of the transistor budget these days is taken by microcode that performs these optimizations on the fly, but with architectures such as Itanium you're seeing a move towards compile time ordering. But now I'm getting sidetracked...
Ever looked at the specification of a Playstation [e-scapegames.co.uk] and wondered how on earth developers got it to do what they had it doing by the end of its lifecycle?This is a time-honoured trend in closed-hardware systems. Look at the last generation of games on the SNES. Look at the scene demos being put out for the Amiga in the mid 90's, essentially a decade old hardware platform at that point.
Ever wondered why early Playstation 2 games bit the weenie?
All first generation titles do not harness the full capacity of a machine; I think this is your point. But on open systems developers don't have to optimize anything, thanks to Moore's Law. Maybe it doesn't strictly live up to the "small-is-beautiful" aesthetic, but software development is about optimizing results. Time will be spent where the greatest payoff is, and since performance boosts are a natural consequence of progress more resources are devoted to development.
Yeah, Luxo Jr was a classic. It's amazing to see how far hardware has come; Luxo was rendered in realtime by a GeForce 3 card in a Macworld demo. I'd link up the demo but according to this interview it is not being released :(
You can't use the integrated video and a separate card for dual display unfortunately. I read that in a preview somewhere.
I really don't want to sound like a zealot or anything like that, but I would like to know if the integrated peripherals (sound, video, NIC) will be supported under Linux. It seems like a no-brainer for the video, given NVIDIA's commendable (though closed) Linux drivers. I remember jumping out and buying a TNT2 when they announced Linux support for the card, then waiting a year for the 4.X drivers to finally arrive. Now that I'm finally going to upgrade my computer, the nForce features are attractive, but I would like to know they are supported at the time of purchase instead of having to while away months wishing I'd bought products with existing support.
:)
Really, this isn't a whinge about NVIDIA. Any company that supports Linux to the parity of their Windows releases deserves respect (are you listening ATI? Cirrus Logic? Guillemot? Creative?) After a few initially iffy driver releases, the very latest versions have been pretty much perfect for me. I just want to know if the features of this chipset will be supported off the bat this time. A kernel module for the NIC is probably not a major issue, but a closed sound-card driver will be hard to reconcile with the ALSA project, set to be the new Linux sound standard in 2.5.x. But any support is better than none, as they have proven in the past.
Flame away zealots
I haven't seen it pointed out yet, but GDK/GTK 1.3 have had AA enabled for a while now, so this is very much an interim thing while we wait for the big gnome 2.0 release.
I've tried it out a bit and generally liked it. There are some problems with font sizes in certain applications, where the font is now larger than the widget, but then again this may be due to the changed font preferences required. It takes a bit of fudging the configurations in Debian, and make sure you have a symlink /etc/X11/XF86Config to your /etc/X11/XF86Config-4 if you're running XF4.0, or the config script dies.
The GTK anti-aliasing is still being handled by the FreeType engine, which is IMHO perceptively as good as it gets. But you're begging for the screenshots aren't you? Here are some tiny morsels for you :)
Seeing as they had a huge edge in technology but are now...non-existent...then yes.
And I'm speaking as someone who had an original 1000, with the paw-print mobo, in '86.
Damned right.
That's twisting thet truth significantly. The major reasons cited in the Wired article for the failure of Linux adoption were:
Furthermore, it appears the ScholarNet initiative is not over yet. The current progress has not achieved the penetration desired, but future iterations hold more promise, as hardware compatibility improves and the "seeded" Linux knowledge from the successful installations trickles down.
Anyone interested in the attractions of Linux implementations in developing countries might want to have a look at a paper I wrote for a final year Engineering unit: postscript version. It has some mistakes in it I haven't corrected yet, but I'm open to revising it :)
hmmm...the daily poll is "favourite terrorist organisation", so I'm guessing it's a pisstake.
...it was ported to the PS/2 with the 1.2.1 release a few weeks ago. The port was done by some Japanese coders. The news here is that Sony has donated a dev platform for further development...so we can all hope Mesa is ported sometime :)
Any word when anti-aliased text rendering is going to be incorporated into the *nix versions?
FWIW I've been using linux too, rp-pppoe.
But I'm guessing you don't live in Perth?
This is 100% correct. Not only that, after they changed it, we were told we couldn't change it to something less obvious for a few days while they "worked out a few kinks".
Telstra Bigpond has been unequivocally the WORST internet service I've ever used, of 4 ISP's tried.
As a BigPond ADSL user I have much to be thankful for. Thanks for the two-weeks downtime last month, changing the user agreement on download restrictions after the contract was signed, and forcing me to call every 4 days to reset my account when an authentication error on your end hangs the connection.
But most of all, thanks for leaking the account passwords through poor security and having the foresight to keep the server down right now so I can't change mine.
I believe the low-latency patches give latencies of the order 1-2ms. RTlinux is overkill for this type of problem, and requires specialised hardware support in the form of new drivers for the soundcards. AFAIK, no commonly used cards have RTlinux driver support...and there are no audio apps that use the RTlinux API either.
summer games was a softie compared to hypersports
There's some really good info and ideas at:
http://www.sospubs.co.uk/sos/jan00/articles/noise. htm
The article is aimed at studio use, where background noise is unacceptable.
More clearly, if you actually read the GPL that comes with Linux it has a clause explicitly permitting linking of non-GPL code with the system, which is considered as "normal use".
Corporate licenses of XP do not have the "fingerprint" restriction, which is why corporate copies of the OS and presumably Office will be the most coveted and pirated of all. Regular joes will end up footing the bill.
I can't wait till my ISP contacts me saying "The RIAA wants us to shut you down because you run a Freenet server" my response will be "Um - ok, whats their beef? Oh copyrighted music? Prove that data lives on my server! What? Its encrypted? Wow. Do I have the keys? Maybe - maybe not. I have no idea. Oh yeah - thats right - you aren't liable for criminal traffic going through your network either if you can't identify it? Well neither can I." Click.
I'd be more like "Encrypted copyright materials on my server? That's really perturbing...I tell you what, I'll pay you $10000 to figure out what the material is...". Then it's DMCA - pow!
Heh.
Understood. I wanted to know if any research has been done to determine the discovery time and whether this would make the scheme feasible...although I'm pretty sure no, for the same reason cryptography is secure these days.
:)
Just spouting random thoughts, jeez
A while ago on slashdot.org an article was posted about a prime number that had the "magical" property of being able to be g'unzip'd to the DeCSS source. The number was represented as a (relatively short) arithmetic expression.
My question is, has anyone put any efforts towards seeing if other, larger pieces of data could be represented in this way? Of course the chances of finding these type of numbers are very slim, but they *are* infinitely large spaces, and the compression would be ultra-compact.
Maybe something for a quantum computer to chew on anyway...
Hey, Digi-Paint wasn't written in Assembler. [Digi] was the original HAM mode paint program, and was pretty slow at the time (but still useable). You're thinking of Brilliance, the wicked fast paint package that took, I believe, 5 man years of coding.
Still, DPaint really did cane, although it was slower towards the end. I'm amazed there isn't a paint program like it for Linux.
Dan.