>>IMO Games are one area where the GNU/Open Source Model is unlikely to work.
>Wrong.
This remains to be seen. If I'm wrong, great! Free games to play. >}:) But the outlook now is that we're going to end up with great game concepts, and great games engines, with crappy artwork.
>>...now a mature software area, and with today's hardware we're almost at the point where brute
>>forcing it will be a "good enough" programming strategy.
>Wrong. Copying graphics to the screen one pixel at a time will _always_ suck.
You're absolutely right. Which is why, with today's hardware, you make calls to the hardware drivers to blit for you. AKA "Brute force".
I'm sorry if you feel trolled, my point wasnt that Open Source couldn't produce good games engines. The basic math for games engines and 3d graphics has not really changed in 10 years, and is unlikely to (barring quantuum computers). My point was that, culturally, the OSS model is unlikely to infect the community of games-oriented artists: It's a different culture, with a different mindset.
Go ask an ambitious and talented Graphic Arts type if you can take his/her work and give it away for free, allowing anyone to copy and re-use it as they see fit. Observe the scowl you are most likely to get in reply.
IMO Games are one area where the GNU/Open Source Model is unlikely to work. Game engines, 3D graphics, etc are now a mature software area, and with today's hardware we're almost at the point where brute forcing it will be a "good enough" programming strategy. So games dont really belong to the programmers like they did in the 70s and 80s. This has been true for some time.
Today's games are largely about the graphical arts, soundtrack, 'look & feel' and (sometimes) storyline. Games are now about art. And artists are, for the most part, a greedy and opportunistic breed.... no 'gift culture' for them, thank you very much.
I can imagine an Open Source games culture where artsy young wannbes would use a Free Software game as a springboard for their professional career... or do little one-shots to impress their friends (I'm thinking of the amiga demo culture of the 80s & 90s here). I could also imagine a world in which a number of cheap-but-decent Shareware games were produced for Linux (this model worked well on the Mac in the 90s actually).
But I cant see Open Source games going much further than that.
PS. I also grow weary of ESR as my "tribe's" representative. We are too diverse a group of people for any one (or two) people to really represent our views. Otherwise slashdot wouldn't be as interesting as it is.
One of the more disturbing subtexts of ESR's use of the word 'tribe' is that, in an actual tribe (think "primitive peoples"), there can be no dissent in the modern sense. All members are of the same blood, have identical culture, and have similar life experiences so there is actually not much in the way of deeply opposed points-of-view. And the Tribal Elders run the show, and tolerate no serious opposition.
Given these attributes, it makes perfect sense for ESR to refer to all hackers as his Tribe: In his view, those with any fundamental disagreements are, by definition, not of His Tribe. And therefore not really hackers. Witness phrases such as "...my report of what the open-source culture knows and believes and wants." If you disagree, you Don't Really Matter.
As a self-described anthropologist/linguist ESR cannot be totally ignorant of the implications of this usage of the word.... and one does sometimes get the impression from his essays that this sort of tribal setup would, indeed, be his ideal social structure for Open Source.
Your best bet would be an SMP Alpha. With a better floating point unit than Intel and AMD, and an outright faster CPU, you'd get a lot more done in less time.
I've never heard of Alphas being used in "real" real-time embedded systems. The chip runs too hot, has too high a per-unit cost, and can have reliability problems in harsh environments (again, because it runs so hot).
Same reason x86 is not that popular in the embedded world. I usually see PPC, ARM, and 68k (yes they still use them). In applications where raw number crunching is really a hard requirement, these computations arent done on the main processor at all: Usually a special DSP chip (say one of TIs) would be used.
Then again, we may be talking apples and oranges here. the Slashdef of the term "real time embedded system" seems to mean something very different from the definition used by industry.
In real-time/embedded systems terminology there is an architecture known as "Foreground/Background". Foreground tasks (often interrupt-driven) run within assigned time-slices. Whenever there is no Foreground task running, the Background task gets to use the time until another Foreground task's turn to run has arrived, in which case the Background task is again pre-empted. The Background task itself is never allowed to preempt one of the Foreground tasks.
What the patent describes is nothing new. The RT-linux real-time tasks are the "Foreground" tasks, and the "non real time OS" is just the Background task. I once worked on a legacy system from the 70s (aerospace application) which used this kind of archtecture, so this idea is old, old old.
A tricky twist in terminology does not make a radical new idea.
What saddens me more is that fact that most here seem to approve of the patent, just because it was someone with LinuxCred(tm) who obtained it.
Software patents, as currently abused, are a serious threat to progress in the field. Just because a patent holder promises to be nice to your friends doesnt make this particular patent any less silly, or any more just.
Or has Open Source gone "Four legs good/Two legs bad" already?
Once a technology has reached that point in its lifecycle where domain name deals, buyouts, IPOs and the like are the chief means of progress, that technology has ceased to be interesting from a technological standpoint.
Corollary #1: Slashdot is now a business page. Expect deals like this to be praised without question.
Corollary #2: Prepare for a massive cultural shift this year. The technically inclined will be moving on to something else very soon.
Ironically, an end (or at least a substantial waning) to the dotcom culture could lead to MORE innovation.
When I was a grad student in the early 90s, there was quite a bit of interesting research taking place at my company. Novel computing architectures. Novel algorithms development. It was fairly easy to get a small but workable grant to investigate, and often implement, almost any feasible theoretical idea you could propose--- and this was during a major recession remember!
Then the dotcoms came. And now?
OS Research is virtually a dead field.
Research into any non-Intel processor architecture has virtually stopped. (Except for quantum computing... but notice: no dotcoms there, folks).
"Multiprocessing" now means SMP on NT or Linux.
Language research was stagnant for quite awhile, thanks to Java and C++. (Functional languages seem to be coming back into vogue now however. Again, notice: no dotcoms in sight).
And the general computing profession has basically morphed from a quasi-academic profession (remember, we used to be thought of as applied mathematicians) into something halfway between a plumber and car mechanic (guys who keep servers running and spin CGI).
I've been telling people for 4 years now that the internet boom is the worst thing that could have happened to the computing profession. If the shallow, soulless and stilting dotcom boom is really, really over then I will know in a couple years if I have been right or not.
...why in the 1950s, 1960s, 1970s and even the 1980s you could send all kinds of trash up there and it would just reach the destination at least somewhat all right? Isn't the overall technology level higher now?
Because in the 50s, 60s and 70s, NASA could spend enough money to do these things right.
$150 - $200 million simply may not be enough to guarantee success for a Mars Probe. Engineering for space is enormously complicated and expensive. You need to use much better parts, the system engineering is harder, the software engineer is much, much harder, the testing, the analysis, etc etc are all far beyond what we as a culture seem able (or willing) to pay for.
But these are the years of the Great Navelgazing. Investors will gladly throw billions of dollars at "web startups" that have yet to show a profit, or any significant advancements to technology... but not one penny for the future of the species. In a cultural climate like this, NASA is lucky to get even the few scraps that Congress is begrudgingly willing to throw it.
The fact is that failures like this will continue to happen until something changes.
In the embedded market, if you work is what matters. Consider too, that what has kept Microsoft where they are today is the OS and the supporting software (Office, for example). In the embedded world, the supporting apps don't matter. (Note how WinCE isn't taking over as others thought it would.)
The "embedded world" issues are more complicated than this. Linux could never take over the real-time embedded market for the same reasons WinCE never could: It is not designed for hard real-time applications. It is not a microkernel architecture. Like UNIX, it is process-based rather than thread (or task) based. And any companies who would need to modify the kernel for their applications would be scared off by the infectious nature of the GPL.
Buying Cygnus gives RedHat access to that company's many support tools and compilers. That does not however automatically give them a free ride into the embedded market. In fact, in the real-time world this acquisition may scare some of us off to the proprietary compilers (Metroworks for PPC, for example). I, for one, would be concerned if the now environment-agnostic gcc/egcs compiler suite were optimized for compiling the Linux kernel and apps. What would happen to the cross-compilers I depend on to do my job?
Great desktop OS? Sure. Embedded real-time OS? No way. Not without changing linux so much that it would no longer be recognizable as linux.
A question for you (and this is absolutely not a flame)...do you think Macs are so efficient because you 'have been using them your whole life', or by design?
That's an interesting question. Probably a little of both. There must be more to my perceived efficiency of the Mac than mere familiarity, because I've been forced to use Windows 95/98/NT at work for almost three years now and have not "gotten used to it" yet. Windows is NOT that easy to use, despite the FUD. Unlike the Mac, I just feel Windows getting in my way, rather than helping me. It just does not feel put together right.
Using Linux or Windows, I still do a lot of stuff from the command line simply because I am a much faster typist that I am with a mouse. It's probably an illness, but I like the script files much more than searching through menus and hitting check boxes.
I actually agree with you about the command line, and I am one of those Mac users who hopes that the upcoming Mac OS X, when released, will still allow users access to the BSD command line. Some tasks, like the renaming of multiple files at once, i/o redirection, pipes etc can really only be done well from a command line.
For now, there is always MacPerl.
The interesting thing is that you can see that the two OSes seem to be converging: Linux is moving towards the best-of-both ideal by trying to develop easy-to-use GUIs such as GNOME or KDE... and the Mac is moving towards the very same ideal by replacing its OS core with a unix variant on top of a mach kernel.
One way or the other, I'm pretty sure that I will FINALLY have the OS I've been dreaming about within 2-3 years.
A question for you (and this is absolutely not a flame)...do you think Macs are so efficient because you 'have been using them your whole life', or by design?
That's an interesting question. Probably a little of both. There must be more to my perceived efficiency of the Mac than mere familiarity, because I've been forced to use Windows 95/98/NT at work for almost three years now and have not "gotten used to it" yet.
Professionally, I am primarily a UNIX and VxWorks developer (I have a Linux box on my desk which I use for prototyping) and sysadmin who has used far too many flavors of unix to remain normal. I've also done quite a bit of Macintosh programming.
At home, I have both an iMac and a linux laptop to play with. Guess which one is used more? The iMac. And no I am not dumbing down off-hours... I do most of my home coding projects on the Mac these days. The Mac is just more fun to use, and when it's MY OWN time, the fun is really all that matters.
Is Linux "fun to use"? Yes. As a challenging and powerful environment to keep one's unix chops up to snuff. It's also the most pleasant of all the unices I've seen or used over the years.
Is the Macintosh "fun to use"? Also yes. As a tool that hardly ever gets between me and whatever it is I am trying to do.
I guess what I am trying to say is that linux is fun in the sense that, say, Rubik's Cube or a challenging puzzle is fun. My Mac is fun in the sense that my guitars are fun. There is plenty of room in my life for both.
I have as much "unix cred" as almost anyone, and I still have nothing but respect for the general design of the MacOS. It bothers me when (as happens all too often) I hear perfectly competent and intelligent linux advocates flaming MacOS by reflex.
I'm not really sure why it has to be one or the other. Why not both?
Slashdot Mirror
>Wrong.
This remains to be seen. If I'm wrong, great! Free games to play. >}:) But the outlook now is that we're going to end up with great game concepts, and great games engines, with crappy artwork.
>>...now a mature software area, and with today's hardware we're almost at the point where brute
>>forcing it will be a "good enough" programming strategy.
>Wrong. Copying graphics to the screen one pixel at a time will _always_ suck.
You're absolutely right. Which is why, with today's hardware, you make calls to the hardware drivers to blit for you. AKA "Brute force".
I'm sorry if you feel trolled, my point wasnt that Open Source couldn't produce good games engines. The basic math for games engines and 3d graphics has not really changed in 10 years, and is unlikely to (barring quantuum computers). My point was that, culturally, the OSS model is unlikely to infect the community of games-oriented artists: It's a different culture, with a different mindset.
Go ask an ambitious and talented Graphic Arts type if you can take his/her work and give it away for free, allowing anyone to copy and re-use it as they see fit. Observe the scowl you are most likely to get in reply.
:Michael
Today's games are largely about the graphical arts, soundtrack, 'look & feel' and (sometimes) storyline. Games are now about art. And artists are, for the most part, a greedy and opportunistic breed.... no 'gift culture' for them, thank you very much.
I can imagine an Open Source games culture where artsy young wannbes would use a Free Software game as a springboard for their professional career... or do little one-shots to impress their friends (I'm thinking of the amiga demo culture of the 80s & 90s here). I could also imagine a world in which a number of cheap-but-decent Shareware games were produced for Linux (this model worked well on the Mac in the 90s actually).
But I cant see Open Source games going much further than that.
:Michael
One of the more disturbing subtexts of ESR's use of the word 'tribe' is that, in an actual tribe (think "primitive peoples"), there can be no dissent in the modern sense. All members are of the same blood, have identical culture, and have similar life experiences so there is actually not much in the way of deeply opposed points-of-view. And the Tribal Elders run the show, and tolerate no serious opposition.
Given these attributes, it makes perfect sense for ESR to refer to all hackers as his Tribe: In his view, those with any fundamental disagreements are, by definition, not of His Tribe. And therefore not really hackers. Witness phrases such as "...my report of what the open-source culture knows and believes and wants." If you disagree, you Don't Really Matter.
As a self-described anthropologist/linguist ESR cannot be totally ignorant of the implications of this usage of the word.... and one does sometimes get the impression from his essays that this sort of tribal setup would, indeed, be his ideal social structure for Open Source.
Depressingly enough, we may already be there.
:Michael
>>Actually, I think patenting this type of machine is pretty reasonable.
Not when said research is funded on the public dime. Then it is not 'reasonable' at all, it is corporate theft.
I've never heard of Alphas being used in "real" real-time embedded systems. The chip runs too hot, has too high a per-unit cost, and can have reliability problems in harsh environments (again, because it runs so hot).
Same reason x86 is not that popular in the embedded world. I usually see PPC, ARM, and 68k (yes they still use them). In applications where raw number crunching is really a hard requirement, these computations arent done on the main processor at all: Usually a special DSP chip (say one of TIs) would be used.
Then again, we may be talking apples and oranges here. the Slashdef of the term "real time embedded system" seems to mean something very different from the definition used by industry.
:Michael (embedded systems/aerospace geek).
LOTS of prior art here.
In real-time/embedded systems terminology there is an architecture known as "Foreground/Background". Foreground tasks (often interrupt-driven) run within assigned time-slices. Whenever there is no Foreground task running, the Background task gets to use the time until another Foreground task's turn to run has arrived, in which case the Background task is again pre-empted. The Background task itself is never allowed to preempt one of the Foreground tasks.
What the patent describes is nothing new. The RT-linux real-time tasks are the "Foreground" tasks, and the "non real time OS" is just the Background task. I once worked on a legacy system from the 70s (aerospace application) which used this kind of archtecture, so this idea is old, old old.
A tricky twist in terminology does not make a radical new idea.
What saddens me more is that fact that most here seem to approve of the patent, just because it was someone with LinuxCred(tm) who obtained it.
Software patents, as currently abused, are a serious threat to progress in the field. Just because a patent holder promises to be nice to your friends doesnt make this particular patent any less silly, or any more just.
Or has Open Source gone "Four legs good/Two legs bad" already?
:Michael
Once a technology has reached that point in its lifecycle where domain name deals, buyouts, IPOs and the like are the chief means of progress, that technology has ceased to be interesting from a technological standpoint.
Corollary #1: Slashdot is now a business page. Expect deals like this to be praised without question.
Corollary #2: Prepare for a massive cultural shift this year. The technically inclined will be moving on to something else very soon.
:Michael
Ironically, an end (or at least a substantial waning) to the dotcom culture could lead to MORE innovation.
When I was a grad student in the early 90s, there was quite a bit of interesting research taking place at my company. Novel computing architectures. Novel algorithms development. It was fairly easy to get a small but workable grant to investigate, and often implement, almost any feasible theoretical idea you could propose--- and this was during a major recession remember!
Then the dotcoms came. And now?
OS Research is virtually a dead field.
Research into any non-Intel processor architecture has virtually stopped. (Except for quantum computing... but notice: no dotcoms there, folks).
"Multiprocessing" now means SMP on NT or Linux.
Language research was stagnant for quite awhile, thanks to Java and C++. (Functional languages seem to be coming back into vogue now however. Again, notice: no dotcoms in sight).
And the general computing profession has basically morphed from a quasi-academic profession (remember, we used to be thought of as applied mathematicians) into something halfway between a plumber and car mechanic (guys who keep servers running and spin CGI).
I've been telling people for 4 years now that the internet boom is the worst thing that could have happened to the computing profession. If the shallow, soulless and stilting dotcom boom is really, really over then I will know in a couple years if I have been right or not.
:M
Because in the 50s, 60s and 70s, NASA could spend enough money to do these things right.
$150 - $200 million simply may not be enough to guarantee success for a Mars Probe. Engineering for space is enormously complicated and expensive. You need to use much better parts, the system engineering is harder, the software engineer is much, much harder, the testing, the analysis, etc etc are all far beyond what we as a culture seem able (or willing) to pay for.
But these are the years of the Great Navelgazing. Investors will gladly throw billions of dollars at "web startups" that have yet to show a profit, or any significant advancements to technology... but not one penny for the future of the species. In a cultural climate like this, NASA is lucky to get even the few scraps that Congress is begrudgingly willing to throw it.
The fact is that failures like this will continue to happen until something changes.
In the embedded market, if you work is what matters. Consider too, that what has kept Microsoft where they are today is the OS and the supporting software (Office, for example). In the embedded world, the supporting apps don't matter. (Note how WinCE isn't taking over as others thought it would.)
The "embedded world" issues are more complicated than this. Linux could never take over the real-time embedded market for the same reasons WinCE never could: It is not designed for hard real-time applications. It is not a microkernel architecture. Like UNIX, it is process-based rather than thread (or task) based. And any companies who would need to modify the kernel for their applications would be scared off by the infectious nature of the GPL.
Buying Cygnus gives RedHat access to that company's many support tools and compilers. That does not however automatically give them a free ride into the embedded market. In fact, in the real-time world this acquisition may scare some of us off to the proprietary compilers (Metroworks for PPC, for example). I, for one, would be concerned if the now environment-agnostic gcc/egcs compiler suite were optimized for compiling the Linux kernel and apps. What would happen to the cross-compilers I depend on to do my job?
Great desktop OS? Sure. Embedded real-time OS? No way. Not without changing linux so much that it would no longer be recognizable as linux.
That's an interesting question. Probably a little of both. There must be more to my perceived efficiency of the Mac than mere familiarity, because I've been forced to use Windows 95/98/NT at work for almost three years now and have not "gotten used to it" yet. Windows is NOT that easy to use, despite the FUD. Unlike the Mac, I just feel Windows getting in my way, rather than helping me. It just does not feel put together right.
Using Linux or Windows, I still do a lot of stuff from the command line simply because I am a much faster typist that I am with a mouse. It's probably an illness, but I like the script files much more than searching through menus and hitting check boxes.
I actually agree with you about the command line, and I am one of those Mac users who hopes that the upcoming Mac OS X, when released, will still allow users access to the BSD command line. Some tasks, like the renaming of multiple files at once, i/o redirection, pipes etc can really only be done well from a command line.
For now, there is always MacPerl.
The interesting thing is that you can see that the two OSes seem to be converging: Linux is moving towards the best-of-both ideal by trying to develop easy-to-use GUIs such as GNOME or KDE... and the Mac is moving towards the very same ideal by replacing its OS core with a unix variant on top of a mach kernel.
One way or the other, I'm pretty sure that I will FINALLY have the OS I've been dreaming about within 2-3 years.
:Michael
That's an interesting question. Probably a little of both. There must be more to my perceived efficiency of the Mac than mere familiarity, because I've been forced to use Windows 95/98/NT at work for almost three years now and have not "gotten used to it" yet.
Professionally, I am primarily a UNIX and VxWorks developer (I have a Linux box on my desk which I use for prototyping) and sysadmin who has used far too many flavors of unix to remain normal. I've also done quite a bit of Macintosh programming.
At home, I have both an iMac and a linux laptop to play with. Guess which one is used more? The iMac. And no I am not dumbing down off-hours... I do most of my home coding projects on the Mac these days. The Mac is just more fun to use, and when it's MY OWN time, the fun is really all that matters.
Is Linux "fun to use"? Yes. As a challenging and powerful environment to keep one's unix chops up to snuff. It's also the most pleasant of all the unices I've seen or used over the years.
Is the Macintosh "fun to use"? Also yes. As a tool that hardly ever gets between me and whatever it is I am trying to do.
I guess what I am trying to say is that linux is fun in the sense that, say, Rubik's Cube or a challenging puzzle is fun. My Mac is fun in the sense that my guitars are fun. There is plenty of room in my life for both.
I have as much "unix cred" as almost anyone, and I still have nothing but respect for the general design of the MacOS. It bothers me when (as happens all too often) I hear perfectly competent and intelligent linux advocates flaming MacOS by reflex.
I'm not really sure why it has to be one or the other. Why not both?
:Michael
...since when are set-top boxes for the consumer market "embedded systems"?
:M