First of all, you're nitpicking over the word "homopobia". The Merriam-Webster On-Line Dictionary gives this definition [...]
Now look up the word "hacker" in your dictionary, use the definition that's there on slashdot and see how far it gets you.:-)
My problem is that the widening of the term "homophobia" to refer to any suggestion that homosexuality might be different from heterosexuality (apart from the obvious difference) desensitises people to real homophobia. There are still cases of homosexual people being discriminated against, assaulted (both physically and verbally) or killed because of their sexual orientation. This kind of violence is almost always caused by irrational fear. That fear is homophobia. Calling what namespace said "homophobia" makes it that much more likely that I will tune out when you say the word "homophobia" next time.
Don't get me wrong. I think namespace was wrong. But he was sincerely wrong, not wrong out of irrational fear.
I don't think you're a homophobe for saying this, but I do believe that you're mistaken.
There are, I think, two fallacies in your argument. One you acknowledge:
I'm willing to concede that homosexual couples can be stable and caring --- so don't bring up the "better them than the abusive/f*d up straight people." Yes, that's true: much better for a kid to be raised outside of the traditional/natural model than abused. [...] All things being equal, it's probably better to be raised by straight parents.
All things are not equal. Every couple is different and until some serious long-term research is done on this, you can't really make generalisations of this kind.
The other fallacy is on the topic of role models. Every boy needs a good male role model and every girl needs a good female role model, this is true. The opposite is also true: research seems to show that heterosexual men, on the whole, get the idea of what a "life partner" should be like from their mothers.
The fallacy is that the role models have to be parents. I know I'm influenced, for example, more by my grandfather than my father. For others it might be an aunt or uncle or a close family friend. Is there any real difference in the male role model arrangement between a lesbian couple raising a child and a heterosexual couple raising a child where the father is never home?
Some about the gay male mind that makes sense of digitals patterns more easily?
As someone else said, it's probably more that geeks find it easier to "come out". Hell, most of us were bullied or worse at school over interest in science or lack of interest in sport. A little thing like what consenting adults do in private is hardly going to matter to us.
As another example, look at religion. There are a number of high-profile Christians who were responsible for computing infrastructure which we all enjoy. Don Knuth and Larry Wall spring to mind immediately; there are quite a few more, too. One might expect in a highly rational field where atheism, agnosticism and non-traditional religions (e.g. Wicca, Discordia, Church of the Subgenius) abound such as this one, especially where people can get extremely emotional (geek rants are the best rants, no doubt about it), that more traditional religion would be an object of ridicule. Not so. It seems that real geeks don't mind what you believe so long as you believe it for the right reasons. (A geek would probably find sheep-like following of religion just as exasperating as sheep-like atheism.)
While Coda did withdraw support (though there is an Open Source project which is an ETF to NIFF converter), it is supported by Mark of the Unicorn, Musitek, Musicware (which probably covers most of the non-Finale market), plus some of the most prolific publishers such as Boosey & Hawkes and Hal Leonard.
If NIFF is a flexible as it claims to be, then there's absolutely no reason why a perfect converter couldn't be written to take lilypond input and spit out a NIFF file (and vice versa).
Actually, that's not a bad idea. Take Lilypond as one of the accepted submission languages, but maintain everything in the repository in one of the more standard formats.
I'm not sure how perfectly convertable they are, though. Converting from NIFF to Lilypond would be lossy at best. Converting the other way is probably like converting from Fortran to C. (Have you ever looked at the output from f2c?)
Naturally, they had to pick the least compatible music notation format there is.
The most compatible is NIFF. It's non-proprietary, and supported by all major music notation software. The only problem is it's a binary format, but then so is PDF. A good cross-platform ASCII format would be SDML, which based on SGML and currently in ISO draft standard status.
After years of pleading with others to stick to published standards, is it too much to hope that we do the same?
Am I the only one who thinks 70 years is a ridiculous amount of time for a dead person to hold onto a copyright?
It gets worse. We geeks often think we're breaking new ground with our hatred of insane copyright law and regulations, but it's been going on for years.
In Australia, performing Grand Right Works (basically anything intended to be performed on stage, such as opera, musical, revue, pantomime or choral work over 20 minutes long) requires paying money to AMCOS, no matter how long the composer has been dead. You heard it right. Performing a Bach Chorale requires paying the copright meisters even though Bach has been dead for 250 years. How can they do this? Simple: if you don't pay them money for these works that are out of copyright, you don't get the rights to copy or perform anything that is under copyright. They have you by the proverbials.
My mother is a music teacher. Music teachers need to bang their collective heads against these ridiculous regulations all the time, because one thing that music teachers do a lot of is get students to perform music, which requires obtaining performing rights, photocopying rights and so forth. She lived in almost perpetual fear of the AMCOS inspectors. At one point (she doesn't do this any more BTW), she kept her cache of photocopied music in the boot (note for Americans: trunk) of her car and only brought into the office that which was needed for that day.
Oh, and just being out of copyright doesn't necessarily help you. I don't know if you've ever read a composer's autograph, but they're often almost illegible. (I know, I edited a Bach's "Musical Offering" once. I should type it up and submit it to Mutopia.) You really need an edited and published version. But editing and publishing slaps a new copyright on that edition.
IMO, if we geeks spoke to musicians and music teachers over the insanity of copyright law, we would find a strong and vocal ally.
If you're an independent artist (whether you've sold any works or not), petition or sue the government for your cut of the money. After all, this is to compensate artists, not labels.
It doesn't work, it's just intact. There's actually an archival issue here. Do we keep CSIRAC "as it was", or do we restore it to working (and keep replacing valves as they burn out at the rate of at least one per day)?
When I was at the University of Melbourne, we were lucky enough to get a guided tour by one of the original members of the computational machinery laboratory. It's quite easy to see how the meme of the computer as ominous "electronic brain" took hold when you can literally walk through it.
CSIRAC not only had a hard disk (one platter, with a motor which delivered such low torque that you needed to put some pressure on the drive belt with a screwdriver to get it to spin up; I believe one of the engineers still has the screwdriver), but it also had a high level language, the interpreter for which fit in its (off the top of my head) 768 words of memory.
Oh, and another anecdote: When CSIRAC lived in The University of Melbourne, it was first housed next to the particle physics laboratory, which caused some scheduling problems, because CSIRAC wouldn't work when the cyclotron was firing. They also had difficulty with the mercury memory in hot weather, but I suspect all the early computers had that problem.
Whether or not you think it's right, it's still censorship. Someone is taking it into their own hands to block us from seeing something they don't think we should.
That couldn't be further from the truth. What's happening is that I decide that someone else has sufficiently similar opinions to mine about what I don't want to see, and ask them to do it for me. That's not censorship, it's outsourcing. Well, I guess you could always call it "self-censorship by proxy" or something equally convoluted.
Now if the RBL isn't what it advertises to be, that's a different question. But that wouldn't be censorship either, it'd be false advertising.
Regardless of what Win2K does, I have a problem with this idea of impeding technological progress by placing arbitrary limits on something like kernel size.
I don't mean to flame you, but you sound like a PL/I or APL programmer thirty years ago. "I have a problem with this idea of impeding technological progress by placing arbitrary limits on the number of keywords a programming language should have."
Bigger ain't necessarily better, as every C programmer knows.
Back in 1985, you would have presumably been criticising any OS that took up more than say 100K of memory, [...]
I still am. Seriously. 100kb is still too big, except possibly in the case of distributed operating systems where you usually need networking in the kernel otherwise the paradigm isn't practical. (More research should remove this requirement, though.)
The flaw in your thinking is the fallacy that more "technological progress" in operating systems requires more non-pagable kernel memory. Real technological progress requires working smarter, not harder.
If I were you I'd report this to the Australian Competition and Consumer Commission. As anyone who keeps up with the news knows, they have some teeth, and if there's any suggestion of a cartel between movie producers and DVD manufacturers (using the CSS licence), I'm sure they'll want to investigate it. In particular, suggest that they might be using cases in other countries to muscle the open source community out of the DVD player business and artifically fixing the price of DVDs using the region coding system.
Pixar's patent on scalar fields on subdivision surfaces has already been granted. This includes some of the techniques used in their cloth animation system (e.g. the "starch" scalar field applied over Geri's shirt to specify how malleable the cloth is over its surface).
Pixar has a history of software patents. Probably the most silly one is a patent they got during the 80s, and extended in the early 90s, on certain kinds of Monte Carlo integration. Oh, yes, and Pixar also claim an interface copyright on the RenderMan API and RIB protocol. They do NOT play nice with so-called "intellectual property".
Pixar would be one of the coolest companies in the world if Steve Jobs weren't in charge.:-)
Maybe, maybe not. If it weren't for mathematicians, we wouldn't have Arrow's Impossibility Theorem. (Do a web search for it if you don't know what I mean.)
I'm willing to believe that Natapoff's argument is actually quite correct, for what it proves.
Here are some assumptions that his model included that I'd like to see removed for a true analysis:
His analysis only covered two-horse races. This is note a bad assumption for town-hall meetings or referendums, but is utterly for election of officials.
Real voters do have lopsided preferences, but they do not always vote for their conscience, because of the fear of vote splitting in non-monotonic voting systems (e.g. plurality or instant runoff). Note that this is a consequence of three-plus candidate elections, because there is no incentive to vote strategically in a two-candidate race.
The other assumption, of course, is that in real elections (except in places like Australia where voting is compulsory), not everyone turns up, and a significant proportion of those who do not turn up (possibly not
The two candidate assumption is not a bad one to start off the analysis, because it's a lot harder to analyse three-plus candidate elections. Preferences can be quite complex and "the will of the people" can be inconsistent.
The main problem, IMO, with the electoral college system is precisely this one: non-major parties are discriminated against. In this election, something like 3.6% of the population did not vote for a major candidate. I'm certain that a lot more wanted to but feared a split. Let's call it 5% since that's a not unreasonable round number. The highly dissenting opinions of one in every twenty people in the USA, not confined to a few areas (which is one of the issues raised in Natapoff's analysis, that of intense lobbying or "vote stacking" in some district tipping the election), but across the board, were effectively ignored by being completely unrepresented amongst the electors.
Natapoff's analysis, while valuable, should definitely not be taken as the last word on the subject.
My supervisor ("adviser" if you're American) always used to say that if there's no source code, it's not science.
It is that simple. You should not be allowed to publish any experimental results, be it benchmarks or what have you, without also releasing enough information for someone else to reproduce your results exactly. For anything nontrivial, that means releasing the source code. If you don't, the experiment is not reproducible, and if it's not reproducible, it's not science.
Yes, I feel strongly about this issue.:-)
BTW, the "nontrivial" disclaimer is important. If you're analysing an algorithm, while it would be courtesy to release a working implementation, pseudocode or enough English to allow a good programmer to reproduce it is of course sufficient. But, for example, I was at a conference some years ago where a guy from Microsoft Research presented a paper on removing priority inversion from Windows NT. Any experimental data from this research is not science without releasing the source to NT, because otherwise there is no way to peer review the data. After all, maybe their technique, when implemented, actually had the effect of avoiding another completely unrelated performance bug.
please stop using racist terms like "black" in this forum
I'm not American, but I grew up in the era where, in the US, the slogans were things like "black power" and "black is beautiful". The word "black" used to be a badge of pride and dignity. When did those people (whatever I'm supposed to call them) decide that "no, we don't want to apply this term to ourselves any more, we now think it's racist"? Or is it only white people who think the word is wrong?
The question is serious, not rhetorical.
Programmers need what to be productive?
on
Is UNIX An OS?
·
· Score: 2
Something that occurred to me...
An operating system is the software that comes with a computer (or OS distribution) that programmers and users need to make themselves productive.
Using this definition at face value, anything that comes without a programming language (since this is the least that a programmer needs to be productive) is not an OS. This rules out Windows 9X, Windows NT, Solaris and Mac OS from being an operating system (unless you count the command interpreter as a programming language).
HURD is THE future, I mean that this tecnhology is the future, at some point Linux itself will be microkernelized or die in the "big-iron" scenery.
perlmonky wrote:
Microkernel is technically more "correct". Macrokernel WORKS.
If you calm down a bit, you'll see that these statements are entirely compatible. Linux technology is the present. I don't think anyone here seriously doubts that. I also don't think that anyone seriously doubts that the open source movement should be producing good quality software for the present, like Linux. But it's ridiculous to think that Linux in its present form will serve the industry forever, any more than the great OSes of the past did (e.g. VMS, which is finally being phased out as we speak).
We (and I mean both the computer industry and the open source movement) should be looking to the future as well as producing good code for the present. There is a more than twenty year gap between Unix and Linux. If we want to avoid the similar gap in the future, we need to be thinking about the future now, as well as we think about the present. That's why statements like "this technology is the future" need to be said.
Think of Linux as the Empire and Hurd as the Foundation, if that helps.:-)
Re:What's wrong with both the HURD and Linux.
on
HURD For 'Big Iron'?
·
· Score: 2
The HURD is based upon the MicroKernel technology which was in vogue in the Computer Science community about 10 years ago. Time has moved on, people have taken the ideas from this technology and incorporated the best bits into the semi-monolithic kernels of today, Microkernels are looked upon as a relic.
Au contraire. Pretty much every new OS of the past ten years uses a microkernel. Look at BeOS, QNX, JavaOS (built on top of Chorus), Windows NT when it was still relatively stable (i.e. pre-version 4.0), Windows CE and Palm OS.
You're thinking of the microkernels of ten years ago which weren't very "micro" (e.g. Windows NT and Mach).
The problem with this approach is that in the real worl this deisng has a massive performance hit.
It's not nearly as massive as you might think, and you can win it back in other ways.
There are a lot of things in a modern OS which count as a "massive performance hit". Device drivers, for example, are there to abstract away hardware details. It would be faster if we didn't use device drivers at all and just talked to the raw hardware, because we'd eliminate a layer of glue code. But we think this is worth the price because the layers above it can be coded more simply and thus with fewer bugs. You lose in one place and win it back in another place.
Virtual memory incurs a massive performance hit, too. But you win it back because user programs are freed from the responsibility of managing what is in core and what isn't. Simpler algorithms, more efficient algorithms, fewer bugs. You win the performance hit back.
Incidentally, the same argument comes up occasionally about garbage collection, but that's another thread.
There is also the problem of the "one size fits all" mentality.. it doesn't work!
Absolutely. That's why modern microkernels work. You can mix and match different bits to make the OS that you want.
Remember that this is the same NSA who produced the original SHA, then made a small modification (SHA-1) to secure it against some attack which they wouldn't tell us about, and nobody outside the NSA has yet been able to work out what the problem actually was.
What gets me is that to qualify for AES, you had to provide (virtual) reams of documentation about the creators' attempts to break it, and with good reason. What did the NSA provide as evidence that their new hashes are secure?
"NET system means that you can mix and match languages without the associated performance impact."
Only because even the languages that are currently used because they're fast when compiled (Ada95, C, C++) will be reduced to interpreted p-code just like everyone else.
The.NET virtual machine is designed to be JIT-compiled. I would be very surprised if anyone is expected to use interpreted code, in fact.
Microsoft has done quite a bit of research on JIT compilation recently and have come up with some pretty cool technology, such as the GBURG code generator-generator, which produces JIT code generators less than 8kb in size and produces code which runs within a factor of 2-4x that of typical compiler-generated code including the time taken to do the JIT translation. Clearly interpretation is not required under.NET.
Slashdot Mirror
Now look up the word "hacker" in your dictionary, use the definition that's there on slashdot and see how far it gets you. :-)
My problem is that the widening of the term "homophobia" to refer to any suggestion that homosexuality might be different from heterosexuality (apart from the obvious difference) desensitises people to real homophobia. There are still cases of homosexual people being discriminated against, assaulted (both physically and verbally) or killed because of their sexual orientation. This kind of violence is almost always caused by irrational fear. That fear is homophobia. Calling what namespace said "homophobia" makes it that much more likely that I will tune out when you say the word "homophobia" next time.
Don't get me wrong. I think namespace was wrong. But he was sincerely wrong, not wrong out of irrational fear.
I don't think you're a homophobe for saying this, but I do believe that you're mistaken.
There are, I think, two fallacies in your argument. One you acknowledge:
All things are not equal. Every couple is different and until some serious long-term research is done on this, you can't really make generalisations of this kind.
The other fallacy is on the topic of role models. Every boy needs a good male role model and every girl needs a good female role model, this is true. The opposite is also true: research seems to show that heterosexual men, on the whole, get the idea of what a "life partner" should be like from their mothers.
The fallacy is that the role models have to be parents. I know I'm influenced, for example, more by my grandfather than my father. For others it might be an aunt or uncle or a close family friend. Is there any real difference in the male role model arrangement between a lesbian couple raising a child and a heterosexual couple raising a child where the father is never home?
As someone else said, it's probably more that geeks find it easier to "come out". Hell, most of us were bullied or worse at school over interest in science or lack of interest in sport. A little thing like what consenting adults do in private is hardly going to matter to us.
As another example, look at religion. There are a number of high-profile Christians who were responsible for computing infrastructure which we all enjoy. Don Knuth and Larry Wall spring to mind immediately; there are quite a few more, too. One might expect in a highly rational field where atheism, agnosticism and non-traditional religions (e.g. Wicca, Discordia, Church of the Subgenius) abound such as this one, especially where people can get extremely emotional (geek rants are the best rants, no doubt about it), that more traditional religion would be an object of ridicule. Not so. It seems that real geeks don't mind what you believe so long as you believe it for the right reasons. (A geek would probably find sheep-like following of religion just as exasperating as sheep-like atheism.)
The geek community is a wonderful thing.
While Coda did withdraw support (though there is an Open Source project which is an ETF to NIFF converter), it is supported by Mark of the Unicorn, Musitek, Musicware (which probably covers most of the non-Finale market), plus some of the most prolific publishers such as Boosey & Hawkes and Hal Leonard.
Actually, that's not a bad idea. Take Lilypond as one of the accepted submission languages, but maintain everything in the repository in one of the more standard formats.
I'm not sure how perfectly convertable they are, though. Converting from NIFF to Lilypond would be lossy at best. Converting the other way is probably like converting from Fortran to C. (Have you ever looked at the output from f2c?)
Naturally, they had to pick the least compatible music notation format there is.
The most compatible is NIFF. It's non-proprietary, and supported by all major music notation software. The only problem is it's a binary format, but then so is PDF. A good cross-platform ASCII format would be SDML, which based on SGML and currently in ISO draft standard status.
After years of pleading with others to stick to published standards, is it too much to hope that we do the same?
It gets worse. We geeks often think we're breaking new ground with our hatred of insane copyright law and regulations, but it's been going on for years.
In Australia, performing Grand Right Works (basically anything intended to be performed on stage, such as opera, musical, revue, pantomime or choral work over 20 minutes long) requires paying money to AMCOS, no matter how long the composer has been dead. You heard it right. Performing a Bach Chorale requires paying the copright meisters even though Bach has been dead for 250 years. How can they do this? Simple: if you don't pay them money for these works that are out of copyright, you don't get the rights to copy or perform anything that is under copyright. They have you by the proverbials.
My mother is a music teacher. Music teachers need to bang their collective heads against these ridiculous regulations all the time, because one thing that music teachers do a lot of is get students to perform music, which requires obtaining performing rights, photocopying rights and so forth. She lived in almost perpetual fear of the AMCOS inspectors. At one point (she doesn't do this any more BTW), she kept her cache of photocopied music in the boot (note for Americans: trunk) of her car and only brought into the office that which was needed for that day.
Oh, and just being out of copyright doesn't necessarily help you. I don't know if you've ever read a composer's autograph, but they're often almost illegible. (I know, I edited a Bach's "Musical Offering" once. I should type it up and submit it to Mutopia.) You really need an edited and published version. But editing and publishing slaps a new copyright on that edition.
IMO, if we geeks spoke to musicians and music teachers over the insanity of copyright law, we would find a strong and vocal ally.
Ah, but does the internet coffee machine comply with the relevant standards?
Actually, that's not true. The speed of light in a vacuum is exactly 299,792,458 metres per second by decree.
Unfortunately nobody remembered to make the length of a metre final.
If you're an independent artist (whether you've sold any works or not), petition or sue the government for your cut of the money. After all, this is to compensate artists, not labels.
It has, and I believe it ran something like three orders of magnitude faster than the original machine on 1995 hardware.
It doesn't work, it's just intact. There's actually an archival issue here. Do we keep CSIRAC "as it was", or do we restore it to working (and keep replacing valves as they burn out at the rate of at least one per day)?
When I was at the University of Melbourne, we were lucky enough to get a guided tour by one of the original members of the computational machinery laboratory. It's quite easy to see how the meme of the computer as ominous "electronic brain" took hold when you can literally walk through it.
CSIRAC not only had a hard disk (one platter, with a motor which delivered such low torque that you needed to put some pressure on the drive belt with a screwdriver to get it to spin up; I believe one of the engineers still has the screwdriver), but it also had a high level language, the interpreter for which fit in its (off the top of my head) 768 words of memory.
Oh, and another anecdote: When CSIRAC lived in The University of Melbourne, it was first housed next to the particle physics laboratory, which caused some scheduling problems, because CSIRAC wouldn't work when the cyclotron was firing. They also had difficulty with the mercury memory in hot weather, but I suspect all the early computers had that problem.
That couldn't be further from the truth. What's happening is that I decide that someone else has sufficiently similar opinions to mine about what I don't want to see, and ask them to do it for me. That's not censorship, it's outsourcing. Well, I guess you could always call it "self-censorship by proxy" or something equally convoluted.
Now if the RBL isn't what it advertises to be, that's a different question. But that wouldn't be censorship either, it'd be false advertising.
I don't mean to flame you, but you sound like a PL/I or APL programmer thirty years ago. "I have a problem with this idea of impeding technological progress by placing arbitrary limits on the number of keywords a programming language should have."
Bigger ain't necessarily better, as every C programmer knows.
I still am. Seriously. 100kb is still too big, except possibly in the case of distributed operating systems where you usually need networking in the kernel otherwise the paradigm isn't practical. (More research should remove this requirement, though.)
The flaw in your thinking is the fallacy that more "technological progress" in operating systems requires more non-pagable kernel memory. Real technological progress requires working smarter, not harder.
Simple. MkLinux is a kernel fork and therefore responsible for the downfall of civilisation. The Hurd, by comparison, isn't.
If I were you I'd report this to the Australian Competition and Consumer Commission. As anyone who keeps up with the news knows, they have some teeth, and if there's any suggestion of a cartel between movie producers and DVD manufacturers (using the CSS licence), I'm sure they'll want to investigate it. In particular, suggest that they might be using cases in other countries to muscle the open source community out of the DVD player business and artifically fixing the price of DVDs using the region coding system.
Let us know how you do.
Pixar's patent on scalar fields on subdivision surfaces has already been granted. This includes some of the techniques used in their cloth animation system (e.g. the "starch" scalar field applied over Geri's shirt to specify how malleable the cloth is over its surface).
Pixar has a history of software patents. Probably the most silly one is a patent they got during the 80s, and extended in the early 90s, on certain kinds of Monte Carlo integration. Oh, yes, and Pixar also claim an interface copyright on the RenderMan API and RIB protocol. They do NOT play nice with so-called "intellectual property".
Pixar would be one of the coolest companies in the world if Steve Jobs weren't in charge. :-)
Maybe, maybe not. If it weren't for mathematicians, we wouldn't have Arrow's Impossibility Theorem. (Do a web search for it if you don't know what I mean.)
I'm willing to believe that Natapoff's argument is actually quite correct, for what it proves.
Here are some assumptions that his model included that I'd like to see removed for a true analysis:
The other assumption, of course, is that in real elections (except in places like Australia where voting is compulsory), not everyone turns up, and a significant proportion of those who do not turn up (possibly not
The two candidate assumption is not a bad one to start off the analysis, because it's a lot harder to analyse three-plus candidate elections. Preferences can be quite complex and "the will of the people" can be inconsistent.
The main problem, IMO, with the electoral college system is precisely this one: non-major parties are discriminated against. In this election, something like 3.6% of the population did not vote for a major candidate. I'm certain that a lot more wanted to but feared a split. Let's call it 5% since that's a not unreasonable round number. The highly dissenting opinions of one in every twenty people in the USA, not confined to a few areas (which is one of the issues raised in Natapoff's analysis, that of intense lobbying or "vote stacking" in some district tipping the election), but across the board, were effectively ignored by being completely unrepresented amongst the electors.
Natapoff's analysis, while valuable, should definitely not be taken as the last word on the subject.
I assume you're doing computer science.
My supervisor ("adviser" if you're American) always used to say that if there's no source code, it's not science.
It is that simple. You should not be allowed to publish any experimental results, be it benchmarks or what have you, without also releasing enough information for someone else to reproduce your results exactly. For anything nontrivial, that means releasing the source code. If you don't, the experiment is not reproducible, and if it's not reproducible, it's not science.
Yes, I feel strongly about this issue. :-)
BTW, the "nontrivial" disclaimer is important. If you're analysing an algorithm, while it would be courtesy to release a working implementation, pseudocode or enough English to allow a good programmer to reproduce it is of course sufficient. But, for example, I was at a conference some years ago where a guy from Microsoft Research presented a paper on removing priority inversion from Windows NT. Any experimental data from this research is not science without releasing the source to NT, because otherwise there is no way to peer review the data. After all, maybe their technique, when implemented, actually had the effect of avoiding another completely unrelated performance bug.
I'm not American, but I grew up in the era where, in the US, the slogans were things like "black power" and "black is beautiful". The word "black" used to be a badge of pride and dignity. When did those people (whatever I'm supposed to call them) decide that "no, we don't want to apply this term to ourselves any more, we now think it's racist"? Or is it only white people who think the word is wrong?
The question is serious, not rhetorical.
Something that occurred to me...
Using this definition at face value, anything that comes without a programming language (since this is the least that a programmer needs to be productive) is not an OS. This rules out Windows 9X, Windows NT, Solaris and Mac OS from being an operating system (unless you count the command interpreter as a programming language).
So there.
LaBola wrote:
perlmonky wrote:
If you calm down a bit, you'll see that these statements are entirely compatible. Linux technology is the present. I don't think anyone here seriously doubts that. I also don't think that anyone seriously doubts that the open source movement should be producing good quality software for the present, like Linux. But it's ridiculous to think that Linux in its present form will serve the industry forever, any more than the great OSes of the past did (e.g. VMS, which is finally being phased out as we speak).
We (and I mean both the computer industry and the open source movement) should be looking to the future as well as producing good code for the present. There is a more than twenty year gap between Unix and Linux. If we want to avoid the similar gap in the future, we need to be thinking about the future now, as well as we think about the present. That's why statements like "this technology is the future" need to be said.
Think of Linux as the Empire and Hurd as the Foundation, if that helps. :-)
Au contraire. Pretty much every new OS of the past ten years uses a microkernel. Look at BeOS, QNX, JavaOS (built on top of Chorus), Windows NT when it was still relatively stable (i.e. pre-version 4.0), Windows CE and Palm OS.
You're thinking of the microkernels of ten years ago which weren't very "micro" (e.g. Windows NT and Mach).
It's not nearly as massive as you might think, and you can win it back in other ways.
There are a lot of things in a modern OS which count as a "massive performance hit". Device drivers, for example, are there to abstract away hardware details. It would be faster if we didn't use device drivers at all and just talked to the raw hardware, because we'd eliminate a layer of glue code. But we think this is worth the price because the layers above it can be coded more simply and thus with fewer bugs. You lose in one place and win it back in another place.
Virtual memory incurs a massive performance hit, too. But you win it back because user programs are freed from the responsibility of managing what is in core and what isn't. Simpler algorithms, more efficient algorithms, fewer bugs. You win the performance hit back.
Incidentally, the same argument comes up occasionally about garbage collection, but that's another thread.
Absolutely. That's why modern microkernels work. You can mix and match different bits to make the OS that you want.
Remember that this is the same NSA who produced the original SHA, then made a small modification (SHA-1) to secure it against some attack which they wouldn't tell us about, and nobody outside the NSA has yet been able to work out what the problem actually was.
What gets me is that to qualify for AES, you had to provide (virtual) reams of documentation about the creators' attempts to break it, and with good reason. What did the NSA provide as evidence that their new hashes are secure?
The .NET virtual machine is designed to be JIT-compiled. I would be very surprised if anyone is expected to use interpreted code, in fact.
Microsoft has done quite a bit of research on JIT compilation recently and have come up with some pretty cool technology, such as the GBURG code generator-generator, which produces JIT code generators less than 8kb in size and produces code which runs within a factor of 2-4x that of typical compiler-generated code including the time taken to do the JIT translation. Clearly interpretation is not required under .NET.