if you want to do voice browsing you will run into IP issues and your choice will be do something encumbered or don't do it at all.
If that is the case, then we certainly do not want voice browsing as a standard. And companies who might stand to gain something from it because of their expertise or proprietary software in that field, will instead gain nothing because they chose not to give up the right to enforce their patents in this area, causing nothing to be standardized.
In other words, if they want to make their bed that way, then they can lie in it... alone.
Regarding W3C, if you think they are insensitive, try IETF's attitude on intellectual property.
IETF is obviously next in line for the RF treatment. All we need is a clear cut issue, that is, we need many eyeballs watching for a patent-encumbered RFC to make it most of the way through the pipeline. I wonder how often this happens? Not often I suppose, but the next time it does, things should get interesting.
Python byte code is orders of magnitude slower than Java or C# code.
You must try psyco, the specializing compiler for Python. I've benchmarked it, it's amazing, it speeds Python up to within a factor of two of C/C++, and it's seamless.
What next, psychic imprinting? It seems MS just can't win, no matter how they price. If they sell at normal prices, you scream that they are robber barons, being usurious, that they should sell the wares at it true value, which is almost nothing. Yet when they sell at "almost nothing" prices, you still scream that this is "dumping". Yet even them GIVING it away is still evil to you. Man, you can't have it all three ways. Either their prices are too high or too low. Pick one and stick with it.
You are a little confused about how predatory business practices work. The name of the game is to subsidize sales when necessary to undercut your competition, and thus distort the market so you can eventually charge above-market prices, as MS does now.
If you cancel your licence you can be sure that they will audit you and if you don't have oiginal media for each and every piece of software that you have installed you can bet they will slam your but.
Well it seems to me that this is likely to backfire quite seriously. Basically, it gives the school a clear year to plan and evaluate a complete switchover, then provides plenty of motivation for the school to make efforts to ensure that not a single piece of Microsoft software remains in use by the campus, as any remaining Microsoft software would create a risk of copyright infringement.
Of course it only works this way now because they don't need Microsoft Office any more, they can use OpenOffice instead. Maybe the school's datacenter will need to keep a few copies of MS SQL around for legacy application, but even there... SQL, it's not that hard to port. And it will run faster on Postgres as a bonus.
The words "hoist by one's own petard" come to mind.
a) Cut and paste text out of a flash page, say: a phone number? (Yes, I've been facted with that problem more than once.)"
The same way you cut and paste out of non-evil jpgs / pngs / gifs (take your pick).
In other words, you don't know how. Well until you can do that, flash is evil, sorry. Note that very few web designers are stupid enough to place site text in gifs, even though the gif will typically look a lot better. Why? Because it annoys users not to be able to select+copy the text. However, the allure of flash animation is such that any such nicety goes out the window, hence we have a proliferation of pretty animated sites with marginal functionality. This could be fixed, after all, it is possible to pick and manipulate objects out of even 3D animations. But not without coordinated development effort and recognition of the problem. It's the "coordinated" part that I don't see happening with Flash.
"b) Turn off the animation?"
By not downloading the plugin? Surely you won't need it of it's all evil.
And conversely, if I don't download the plugin because animation can't be disabled, then Flash is evil, right?
Anyway, the correct answer to (b) is "Run Mozilla", because Mozilla lets you control Flash animation, even if you have to crawl through a few menus to do it. So we are down to one reason for Flash being evil. I could add more of course, but they would not necessarily be related to basic functionality. Actually, I like the overall concept and I like the compactness, it's a welcome relief from the "size doesn't matter" mantra of the XML mafia. However, given a choice between compactness and the other benefits XML, SVG etc offer, I'll take the latter, thanks. Perhaps Flash does have a role to play in the web ecology, beyond pacification of refugee couch potatoes, or perhaps it will prove to be more like the VESA local bus, a stopgap solution that took over the world for a while, then was replaced by a more carefully thought out standard.
"How about somebody write a Python installer in Java? Let's get something genuinely useful out of this"
Jython [jython.org]s an implementation of the high-level, dynamic, object-oriented language Python written in 100% Pure Java, and seamlessly integrated with the Java platform. It thus allows you to run Python on any Java platform.
Right, I know about that one but I haven't checked how well it works. I was thinking more along the lines of having a small Java program install something a complete Python platform that then runs independently of the Java Platform. There are a few reasons for wanting this, not the least of which is that the Python tool chain is completely open. Strange as it may seem, I'm also interest in Python for performance reasons, ever since I tried this Python specializing compiler. Before that I had my doubts about python, because I'd measured slowdowns of up to several hundred times vs optimized C. Now, with this work by Armin Rigo I feel Python is really looking like the tool that can do it all, as much as any tool can. With a fraction of the time and money put into it, Python+Psyco is now running within striking distance of Hotspot, which is in itself an impressive performer.
So I'd like to see a seamless way for the whole Python toolchain to be installed and then to operate independently of Java, and as I see it, Java is exactly the tool that can automate that, i.e., user sees |"Get Python" button, user clicks, user gets Python, user no longer needs Java for Python-scripted pages.
Java will kill itself as a broswer applet technology on its own.
I have to admit it's been less than awe-inspiring so far, but that is due in large part to Microsoft's manuevering to prevent the obviously needed improvements to the browser-embedded version from happening, essentially freezing web java at 1.0, little grey boxes and all.
That will start to change now, at least it's a chance to show what Java can really do in a browser. I'd like to see a browser-embedded chat client that doesn't suck, for example. I can see a lot of uses on the easy-installer side, for example, you could provide a real hand-holding type of environment for a first-time Linux install over the web, but providing a simple browser-based installer in Java, enough to download a better installer. It hasn't been possible to seriously contemplate nice hacks like that until now, and of course there were the efficiency problems that had to be solved. Which are now pretty much solved, in fact, I was quite amazed when I ran some benchmarks a couple months ago and found Java (Hotspot) was running typical benchmark programs faster than gcc C/C++. Granted, gcc is not the greatest speeddaemon in the world, but still.
Personally, I won't suddenly start doing a whole lot of Java - I just do not like massive heavyweight things, even if today's hardware makes them usable. Then there is the fact that Sun still hasn't let the other shoe drop on the open source side. But all the same, I'm glad that those who do like programming and running Java are going to benefit from this. I suppose I'll be running more Java programs.
How about somebody write a Python installer in Java? Let's get something genuinely useful out of this:-)
"John wants Linux to go off and be totally experimental and new - presumably so he can recapture that excitement of the early years of the PC revolution"
No. He wants it 'to go off...' so he can ridicule it further. Have you ever read what this irrelavent asshole has written?
I wouldn't describe John Dvorak as irrelevent. He's quite widely read, and I'd say, by a lot of influential people. He is also, as nearly as I can tell, his own man and tends to say what he really thinks, regardless of who it ticks off. And he's a generally critical of Microsoft, which you'd know if you followed his columns. He wasn't always that way, but he figured out what was happening earlier than most of the mass of clueless journalists.
Reading between the lines, I can see he's just about ready to make "the switch", and like many switchers, he's idealizing Linux a little too much. He wants it to be not just great, but perfect. I can't say I didn't make the same mistake myself. You get over it.
Like a mainstream political party, Microsoft has firmly occupied the center, as that is way to maximize the allegiance of customers. John wants Linux to go off and be totally experimental and new - presumably so he can recapture that excitement of the early years of the PC revolution - but what happens is, as soon as you move away from center you lose appeal to those who don't like the direction you moved in. So John's recipe for Linux's success is really just a recipe for marginalization.
Another point he's missed so far is that Linux doesn't just move in one direction, it moves in many directions at once, so that you have a number of complete, well-developed environments each of which caters to certain tastes, all the way from text mode consoles to kde, which is more-or-less Linux for windows refugees, to experimental 3D environments. I suppose he would come back with the usual argument about how it doesn't make sense to divide effort across all those different projects, but then he'd just be ignoring one of Linux's great strengths, which is the sheer number of coders involved. In fact, trying to get them working obediently all on the same project at the same time would be shear insanity.
John, if you're reading this, and I guess you will, what you have to realize is that you do get to escape your boring old desktop metaphor and try something different, like a Tivo, which doesn't look like a desktop at all, plus you get to keep working the same way you always did, if that's what you want. It's about choice, and that's what Linux has. How's that for something new?
And, I don't know how you should interpret this, but svn has a handful of paid developers -- arch has none and, in fact, I'm literally days away from homelessness.
There's no way somebody with your talent would not be snapped up, if you're known to be available. But I suppose you're not interested unless you can work on Arch fulltime?
I agree with the other respondent that the dependancy on shell scripts is a perceptual problem. Considered Python? Dependency on diff is not so great either. Actually, the binary diff used by Subversion and XDelta (the code came from XDelta I believe) is only a couple of hundred lines long, and very easy to use.
Subversion impressed our company developers by its speed (subjectively, considerably faster than CVS for comparable operations) and its user-friendlyness. It's the details, stuff like automatic detection of binary files, that makes life for the dev people easier.
Oh yes, and the command syntax is suddenly sensible. For example, to check out subversion using subversion:
svn co http://svn.collab.net/repos/svn/trunk subversion
No CVSROOT, no -dgwana:gwana:gwana, no legacy silliness.
And it has directory versioning, another must. Plus a whole lot of very-nices. For example, the version history is stored in the form of exact binary diffs (xdeltas) whereas cvs uses rcs, which uses a text diff, which is horribly, unspeakably crude, not to mention inefficient in both time and space. As a bonus, Subversion doesn't have to do anything special to handle binary files.
Capitalism is wonderful. It says that if you can sell enough of an item at $25, then that article is worth $25 -- even if it only costs you $2.50 to make and distribute. You are under no obligation to sell the item at a price reflecting the cost, it only reflects the supply of the item, and the demand for the item. If you can find enough people to buy that item at that cost, then that is what it is worth. If it costs you more to make an item than you get back by selling it, you go out of business (or demand government subsidies).
You ignored the fact that natural market forces can be distorted by, for example, monopoly control of promotional distribution channels, which the RIAA enjoys. The corollary to this is that, to maintain this market distortion, alternative channels of promotion and distribution must be prevented from developing. The RIAA, backed by a helpful clutch of lawmakers, has set out methodically to do this, and that is why consumers have every right to complain about the high, monopoly fueled prices of CDs, not to mention the underhanded tactics employed by the RIAA to maintain them.
"Then perhaps you can explain to me what that reasoning is, because I do not understand it."
Sure, look here [slashdot.org].
Err, did you just put forward an unsupported Slashdot post to buttress your argument?
In a nutshell, a lot of the code in the drivers does not belong to nVidia. Instead it was written and contributed or licensed to nVidia under highly restrictive licensing. Therefore nVidia cannot release the source without each of the other parties' explicit permission.
That argument does not hold water. If it were true, and I have reason to believe otherwise (they have a competent in-house driver team) but suppose for the sake of argument it were true, then they could just put those parts in a binary module and open the rest. Since they have not done that, I conclude that contractual restriction is not the reason they don't provide the source.
You've got to give nVidia some credit. They do a hell of a job supporting the *nix community. Better than any other GPU manufacturer ever has -- ATI and 3dfx both included.
That is not true. Both ATI and Matrox provide much better technical documentation and are much more helpful to developers than NVidia. So, they get some credit for having providing drivers at all, but they fail the test of doing what is best for their customers.
So... since ATI is winning both the performance race and the mindshare sweepstakes, don't you think it's about time for NVidia to pull their collective heads out of their butts, before they go the way of 3Dfx?
Your AMD or Intel machine will not get anything close to the mips/watt ratio that the Transmeta does. The LANL people go on to conclude that the Transmeta is cheaper in the long run as well, because of power (including cooling) and space savings. Faster, I'll grant you, but one out of three is pretty poor batting, certainly not worth a moderation of 5.
For a cluster, the faster argument goes out the window as well, because the performance equation comes down to mips/watt, mips/cubic foot and mips/$$$, in all of which Transmeta leads AMD and Intel.
Now I don't know about you, but I find the monolithic, nuclear reactor core kind of box is getting less and less interesting as time goes by, and what I really want is a box full of much more efficient processors, all dirt-cheap of course. I'll admit that that there's no way for the typical home user to get into this kind of system for a price that competes with a single, Athlon or P4, but that's this year. Check again next year.
"I can only think of two compelling reasons for failing to release source code or specifications:
1) Secret contracts or other means of control imposed by a third party interested in impeding the progress of open source software
2) Management ineptitude"
A little from column A, a little from column B.. I don't think it's anyting to do with impeding OSS's progress, more likely that the implementations (i.e. OpenGL) they've licensed for their drivers are under a non GPL compatible license to them.
Even if true (which seems a stretch) how would that stop them from releasing the hardware specs?
but this is posted every time the topic comes up. NVidia can't release the drivers because of legal reasons. There are things in the code that they do not own, thus cannot release.
Sigh. If that was true, they'd put those things in a binary module and open-source the rest. So it's obviously not true.
Companies that are bottom line driven (accept it, you won't change capitalism over night) can't usually afford to jump in head first. This goes double for Nvidia, ATI, etc. Their driver source is like a blueprint of the important parts of their hardware.
That is incorrect. If a competitor, say ATI, is interested in what goes on in NVidia's binary driver they can and will step through it with a debugger or disassemble it, so not providing the driver in source form is at best a minor inconvenience from the competition's point of view.
You might argue that the competition could shorten their driver development time by reading NVidia's driver source. However, since the driver is highly hardware specific, this approach would likely waste more programmer time than it saved, and anyway, could be done with only slightly more effort without the source code. You might argue that part of the rendering algorithm is implemented in the driver, so it should be kept secret from the competition. But that is just the source vs binary argument over again, which doesn't get any more valid through repetition.
In my opinion, failing to release driver source or programming specifications hurts both the hardware vendor and the user. There is no supportable justification for it. Look how much good it did 3Dfx to keep their specs secret, for example.
Conversely, a hardware company that is falling behind in the performance sweepstakes as NVidia is now, only stands to gain by releasing as much information about programming their hardware as they possibly can, so at least the drivers will be the best they possibly can be.
I can only think of two compelling reasons for failing to release source code or specifications:
1) Secret contracts or other means of control imposed by a third party interested in impeding the progress of open source software
I won't criticize NVidia too harshly for distributing binary-only drivers -- I understand their reasoning and I accept it.
Then perhaps you can explain to me what that reasoning is, because I do not understand it. But do not say that it is to preserve their trade secrets, because binary modules are no safer from determined prying eyes than source code is. Nor should you say that it is because they are contractually prohibited from releasing parts of the code they they do not own - if this were true, only those parts would need to be placed in a binary module. That leaves... err, exactly what?
Read the subject heading. None of your comment has anything to do with FilmGIMP. You can provide exceptions to every rule, yes. But, the original poster's claim that increased nVidia support was due to an application that doesn't use any 3D pipeline is baseless and incorrect.
Actually, I'd never heard of FilmGimp before this thread, but now that I've read upon it a little, I have to conclude that you've failed to grasp the importance of this project, besides ignoring the obvious fact that 2D apps play a major part in driving the adoption of 3D apps in this industry. So I'll even pull back on my original guess that NVidia doesn't know or care about FilmGimp. Remember, they have always wanted to break out of the consumer games space into the high-value workstation market. I think they can read the writing on the wall.
Id software can practically drive the industry sometimes, and without decent driver support for the platform Id would have a hard time putting out games like doom III or Quake III for the linux platform.
Perhaps that had something to do with John Carmack joining the Utah GLX project.
"The big push is probably from big studios that use Linux tools such as Film Gimp."
Right, because workstations that use a 2D, time based, painting program need programmable pixel shaders, programmable vertex shaders, hardware transform and lighting, massive fill rate, AGP 8X transfer speeds, and astronomical triangle throughput.
See comments above re potential usefulness of 3D accelerated rendering in 2D graphics, but more immediately, 2D artists need to be able to run 3D animations or renderings for a variety of reasons, for example, to see their work in context (i.e., model skins or painted backdrops) or perhaps they may be provided with 3D models as a starting point from which they develop hand-painted cells.
Slashdot Mirror
if you want to do voice browsing you will run into IP issues and your choice will be do something encumbered or don't do it at all.
If that is the case, then we certainly do not want voice browsing as a standard. And companies who might stand to gain something from it because of their expertise or proprietary software in that field, will instead gain nothing because they chose not to give up the right to enforce their patents in this area, causing nothing to be standardized.
In other words, if they want to make their bed that way, then they can lie in it... alone.
Regarding W3C, if you think they are insensitive, try IETF's attitude on intellectual property.
IETF is obviously next in line for the RF treatment. All we need is a clear cut issue, that is, we need many eyeballs watching for a patent-encumbered RFC to make it most of the way through the pipeline. I wonder how often this happens? Not often I suppose, but the next time it does, things should get interesting.
Python byte code is orders of magnitude slower than Java or C# code.
You must try psyco, the specializing compiler for Python. I've benchmarked it, it's amazing, it speeds Python up to within a factor of two of C/C++, and it's seamless.
"Microsoft's strategy seems to be to extract all the cash from universities that the market will bear, without starting a rebellion."
So they're operating correctly for a company in a free-market environment?
They're operating correctly for a company that has a monopoly and doesn't mind breaking the law.
What next, psychic imprinting? It seems MS just can't win, no matter how they price. If they sell at normal prices, you scream that they are robber barons, being usurious, that they should sell the wares at it true value, which is almost nothing. Yet when they sell at "almost nothing" prices, you still scream that this is "dumping". Yet even them GIVING it away is still evil to you. Man, you can't have it all three ways. Either their prices are too high or too low. Pick one and stick with it.
You are a little confused about how predatory business practices work. The name of the game is to subsidize sales when necessary to undercut your competition, and thus distort the market so you can eventually charge above-market prices, as MS does now.
If you cancel your licence you can be sure that they will audit you and if you don't have oiginal media for each and every piece of software that you have installed you can bet they will slam your but.
Well it seems to me that this is likely to backfire quite seriously. Basically, it gives the school a clear year to plan and evaluate a complete switchover, then provides plenty of motivation for the school to make efforts to ensure that not a single piece of Microsoft software remains in use by the campus, as any remaining Microsoft software would create a risk of copyright infringement.
Of course it only works this way now because they don't need Microsoft Office any more, they can use OpenOffice instead. Maybe the school's datacenter will need to keep a few copies of MS SQL around for legacy application, but even there... SQL, it's not that hard to port. And it will run faster on Postgres as a bonus.
The words "hoist by one's own petard" come to mind.
"If Flash is not evil, then how do you:
a) Cut and paste text out of a flash page, say: a phone number? (Yes, I've been facted with that problem more than once.)"
The same way you cut and paste out of non-evil jpgs / pngs / gifs (take your pick).
In other words, you don't know how. Well until you can do that, flash is evil, sorry. Note that very few web designers are stupid enough to place site text in gifs, even though the gif will typically look a lot better. Why? Because it annoys users not to be able to select+copy the text. However, the allure of flash animation is such that any such nicety goes out the window, hence we have a proliferation of pretty animated sites with marginal functionality. This could be fixed, after all, it is possible to pick and manipulate objects out of even 3D animations. But not without coordinated development effort and recognition of the problem. It's the "coordinated" part that I don't see happening with Flash.
"b) Turn off the animation?"
By not downloading the plugin? Surely you won't need it of it's all evil.
And conversely, if I don't download the plugin because animation can't be disabled, then Flash is evil, right?
Anyway, the correct answer to (b) is "Run Mozilla", because Mozilla lets you control Flash animation, even if you have to crawl through a few menus to do it. So we are down to one reason for Flash being evil. I could add more of course, but they would not necessarily be related to basic functionality. Actually, I like the overall concept and I like the compactness, it's a welcome relief from the "size doesn't matter" mantra of the XML mafia. However, given a choice between compactness and the other benefits XML, SVG etc offer, I'll take the latter, thanks. Perhaps Flash does have a role to play in the web ecology, beyond pacification of refugee couch potatoes, or perhaps it will prove to be more like the VESA local bus, a stopgap solution that took over the world for a while, then was replaced by a more carefully thought out standard.
"How about somebody write a Python installer in Java? Let's get something genuinely useful out of this"
Jython [jython.org]s an implementation of the high-level, dynamic, object-oriented language Python written in 100% Pure Java, and seamlessly integrated with the Java platform. It thus allows you to run Python on any Java platform.
Right, I know about that one but I haven't checked how well it works. I was thinking more along the lines of having a small Java program install something a complete Python platform that then runs independently of the Java Platform. There are a few reasons for wanting this, not the least of which is that the Python tool chain is completely open. Strange as it may seem, I'm also interest in Python for performance reasons, ever since I tried this Python specializing compiler. Before that I had my doubts about python, because I'd measured slowdowns of up to several hundred times vs optimized C. Now, with this work by Armin Rigo I feel Python is really looking like the tool that can do it all, as much as any tool can. With a fraction of the time and money put into it, Python+Psyco is now running within striking distance of Hotspot, which is in itself an impressive performer.
So I'd like to see a seamless way for the whole Python toolchain to be installed and then to operate independently of Java, and as I see it, Java is exactly the tool that can automate that, i.e., user sees |"Get Python" button, user clicks, user gets Python, user no longer needs Java for Python-scripted pages.
Flash is a tool, it is not in itself evil or good.
If Flash is not evil, then how do you:
a) Cut and paste text out of a flash page, say: a phone number? (Yes, I've been facted with that problem more than once.)
b) Turn off the animation?
Java will kill itself as a broswer applet technology on its own.
:-)
I have to admit it's been less than awe-inspiring so far, but that is due in large part to Microsoft's manuevering to prevent the obviously needed improvements to the browser-embedded version from happening, essentially freezing web java at 1.0, little grey boxes and all.
That will start to change now, at least it's a chance to show what Java can really do in a browser. I'd like to see a browser-embedded chat client that doesn't suck, for example. I can see a lot of uses on the easy-installer side, for example, you could provide a real hand-holding type of environment for a first-time Linux install over the web, but providing a simple browser-based installer in Java, enough to download a better installer. It hasn't been possible to seriously contemplate nice hacks like that until now, and of course there were the efficiency problems that had to be solved. Which are now pretty much solved, in fact, I was quite amazed when I ran some benchmarks a couple months ago and found Java (Hotspot) was running typical benchmark programs faster than gcc C/C++. Granted, gcc is not the greatest speeddaemon in the world, but still.
Personally, I won't suddenly start doing a whole lot of Java - I just do not like massive heavyweight things, even if today's hardware makes them usable. Then there is the fact that Sun still hasn't let the other shoe drop on the open source side. But all the same, I'm glad that those who do like programming and running Java are going to benefit from this. I suppose I'll be running more Java programs.
How about somebody write a Python installer in Java? Let's get something genuinely useful out of this
"John wants Linux to go off and be totally experimental and new - presumably so he can recapture that excitement of the early years of the PC revolution"
No. He wants it 'to go off...' so he can ridicule it further. Have you ever read what this irrelavent asshole has written?
I wouldn't describe John Dvorak as irrelevent. He's quite widely read, and I'd say, by a lot of influential people. He is also, as nearly as I can tell, his own man and tends to say what he really thinks, regardless of who it ticks off. And he's a generally critical of Microsoft, which you'd know if you followed his columns. He wasn't always that way, but he figured out what was happening earlier than most of the mass of clueless journalists.
Reading between the lines, I can see he's just about ready to make "the switch", and like many switchers, he's idealizing Linux a little too much. He wants it to be not just great, but perfect. I can't say I didn't make the same mistake myself. You get over it.
Like a mainstream political party, Microsoft has firmly occupied the center, as that is way to maximize the allegiance of customers. John wants Linux to go off and be totally experimental and new - presumably so he can recapture that excitement of the early years of the PC revolution - but what happens is, as soon as you move away from center you lose appeal to those who don't like the direction you moved in. So John's recipe for Linux's success is really just a recipe for marginalization.
Another point he's missed so far is that Linux doesn't just move in one direction, it moves in many directions at once, so that you have a number of complete, well-developed environments each of which caters to certain tastes, all the way from text mode consoles to kde, which is more-or-less Linux for windows refugees, to experimental 3D environments. I suppose he would come back with the usual argument about how it doesn't make sense to divide effort across all those different projects, but then he'd just be ignoring one of Linux's great strengths, which is the sheer number of coders involved. In fact, trying to get them working obediently all on the same project at the same time would be shear insanity.
John, if you're reading this, and I guess you will, what you have to realize is that you do get to escape your boring old desktop metaphor and try something different, like a Tivo, which doesn't look like a desktop at all, plus you get to keep working the same way you always did, if that's what you want. It's about choice, and that's what Linux has. How's that for something new?
And, I don't know how you should interpret this, but svn has a handful of paid developers -- arch has none and, in fact, I'm literally days away from homelessness.
There's no way somebody with your talent would not be snapped up, if you're known to be available. But I suppose you're not interested unless you can work on Arch fulltime?
I agree with the other respondent that the dependancy on shell scripts is a perceptual problem. Considered Python? Dependency on diff is not so great either. Actually, the binary diff used by Subversion and XDelta (the code came from XDelta I believe) is only a couple of hundred lines long, and very easy to use.
Subversion impressed our company developers by its speed (subjectively, considerably faster than CVS for comparable operations) and its user-friendlyness. It's the details, stuff like automatic detection of binary files, that makes life for the dev people easier.
Oh yes, and the command syntax is suddenly sensible. For example, to check out subversion using subversion:
svn co http://svn.collab.net/repos/svn/trunk subversion
No CVSROOT, no -dgwana:gwana:gwana, no legacy silliness.
And it has directory versioning, another must. Plus a whole lot of very-nices. For example, the version history is stored in the form of exact binary diffs (xdeltas) whereas cvs uses rcs, which uses a text diff, which is horribly, unspeakably crude, not to mention inefficient in both time and space. As a bonus, Subversion doesn't have to do anything special to handle binary files.
Capitalism is wonderful. It says that if you can sell enough of an item at $25, then that article is worth $25 -- even if it only costs you $2.50 to make and distribute. You are under no obligation to sell the item at a price reflecting the cost, it only reflects the supply of the item, and the demand for the item. If you can find enough people to buy that item at that cost, then that is what it is worth. If it costs you more to make an item than you get back by selling it, you go out of business (or demand government subsidies).
You ignored the fact that natural market forces can be distorted by, for example, monopoly control of promotional distribution channels, which the RIAA enjoys. The corollary to this is that, to maintain this market distortion, alternative channels of promotion and distribution must be prevented from developing. The RIAA, backed by a helpful clutch of lawmakers, has set out methodically to do this, and that is why consumers have every right to complain about the high, monopoly fueled prices of CDs, not to mention the underhanded tactics employed by the RIAA to maintain them.
"Then perhaps you can explain to me what that reasoning is, because I do not understand it."
Sure, look here [slashdot.org].
Err, did you just put forward an unsupported Slashdot post to buttress your argument?
In a nutshell, a lot of the code in the drivers does not belong to nVidia. Instead it was written and contributed or licensed to nVidia under highly restrictive licensing. Therefore nVidia cannot release the source without each of the other parties' explicit permission.
That argument does not hold water. If it were true, and I have reason to believe otherwise (they have a competent in-house driver team) but suppose for the sake of argument it were true, then they could just put those parts in a binary module and open the rest. Since they have not done that, I conclude that contractual restriction is not the reason they don't provide the source.
You've got to give nVidia some credit. They do a hell of a job supporting the *nix community. Better than any other GPU manufacturer ever has -- ATI and 3dfx both included.
That is not true. Both ATI and Matrox provide much better technical documentation and are much more helpful to developers than NVidia. So, they get some credit for having providing drivers at all, but they fail the test of doing what is best for their customers.
So... since ATI is winning both the performance race and the mindshare sweepstakes, don't you think it's about time for NVidia to pull their collective heads out of their butts, before they go the way of 3Dfx?
Or, build your own Intel- or AMD-based computer that's cheaper, faster, and uses less power.
Not according to the guys who know what they're talking about.
Your AMD or Intel machine will not get anything close to the mips/watt ratio that the Transmeta does. The LANL people go on to conclude that the Transmeta is cheaper in the long run as well, because of power (including cooling) and space savings. Faster, I'll grant you, but one out of three is pretty poor batting, certainly not worth a moderation of 5.
For a cluster, the faster argument goes out the window as well, because the performance equation comes down to mips/watt, mips/cubic foot and mips/$$$, in all of which Transmeta leads AMD and Intel.
Now I don't know about you, but I find the monolithic, nuclear reactor core kind of box is getting less and less interesting as time goes by, and what I really want is a box full of much more efficient processors, all dirt-cheap of course. I'll admit that that there's no way for the typical home user to get into this kind of system for a price that competes with a single, Athlon or P4, but that's this year. Check again next year.
"I can only think of two compelling reasons for failing to release source code or specifications:
1) Secret contracts or other means of control imposed by a third party interested in impeding the progress of open source software
2) Management ineptitude"
A little from column A, a little from column B.. I don't think it's anyting to do with impeding OSS's progress, more likely that the implementations (i.e. OpenGL) they've licensed for their drivers are under a non GPL compatible license to them.
Even if true (which seems a stretch) how would that stop them from releasing the hardware specs?
Not that i'm trying to be an ass,
but this is posted every time the topic comes up. NVidia can't release the drivers because of legal reasons. There are things in the code that they do not own, thus cannot release.
Sigh. If that was true, they'd put those things in a binary module and open-source the rest. So it's obviously not true.
Companies that are bottom line driven (accept it, you won't change capitalism over night) can't usually afford to jump in head first. This goes double for Nvidia, ATI, etc. Their driver source is like a blueprint of the important parts of their hardware.
That is incorrect. If a competitor, say ATI, is interested in what goes on in NVidia's binary driver they can and will step through it with a debugger or disassemble it, so not providing the driver in source form is at best a minor inconvenience from the competition's point of view.
You might argue that the competition could shorten their driver development time by reading NVidia's driver source. However, since the driver is highly hardware specific, this approach would likely waste more programmer time than it saved, and anyway, could be done with only slightly more effort without the source code. You might argue that part of the rendering algorithm is implemented in the driver, so it should be kept secret from the competition. But that is just the source vs binary argument over again, which doesn't get any more valid through repetition.
In my opinion, failing to release driver source or programming specifications hurts both the hardware vendor and the user. There is no supportable justification for it. Look how much good it did 3Dfx to keep their specs secret, for example.
Conversely, a hardware company that is falling behind in the performance sweepstakes as NVidia is now, only stands to gain by releasing as much information about programming their hardware as they possibly can, so at least the drivers will be the best they possibly can be.
I can only think of two compelling reasons for failing to release source code or specifications:
1) Secret contracts or other means of control imposed by a third party interested in impeding the progress of open source software
2) Management ineptitude
Which do you think it is?
I won't criticize NVidia too harshly for distributing binary-only drivers -- I understand their reasoning and I accept it.
Then perhaps you can explain to me what that reasoning is, because I do not understand it. But do not say that it is to preserve their trade secrets, because binary modules are no safer from determined prying eyes than source code is. Nor should you say that it is because they are contractually prohibited from releasing parts of the code they they do not own - if this were true, only those parts would need to be placed in a binary module. That leaves... err, exactly what?
Read the subject heading. None of your comment has anything to do with FilmGIMP. You can provide exceptions to every rule, yes. But, the original poster's claim that increased nVidia support was due to an application that doesn't use any 3D pipeline is baseless and incorrect.
Actually, I'd never heard of FilmGimp before this thread, but now that I've read up on it a little, I have to conclude that you've failed to grasp the importance of this project, besides ignoring the obvious fact that 2D apps play a major part in driving the adoption of 3D apps in this industry. So I'll even pull back on my original guess that NVidia doesn't know or care about FilmGimp. Remember, they have always wanted to break out of the consumer games space into the high-value workstation market. I think they can read the writing on the wall.
Id software can practically drive the industry sometimes, and without decent driver support for the platform Id would have a hard time putting out games like doom III or Quake III for the linux platform.
Perhaps that had something to do with John Carmack joining the Utah GLX project.
"The big push is probably from big studios that use Linux tools such as Film Gimp."
Right, because workstations that use a 2D, time based, painting program need programmable pixel shaders, programmable vertex shaders, hardware transform and lighting, massive fill rate, AGP 8X transfer speeds, and astronomical triangle throughput.
See comments above re potential usefulness of 3D accelerated rendering in 2D graphics, but more immediately, 2D artists need to be able to run 3D animations or renderings for a variety of reasons, for example, to see their work in context (i.e., model skins or painted backdrops) or perhaps they may be provided with 3D models as a starting point from which they develop hand-painted cells.