Yeah, but.... While I love Cryptonomicon it was written well past the time of Turing's life and seemed, at least to me, to have a relatively chirpy view of what it would like to be gay at that time. The Hodges biography is much better about that, although possibly inclined to view too much of Turing's life and work through that lens.
I've got to take issue with you on the claims you make. Sure, patents may have accelerated the desire to hold research closely, but the pressure to be the first to publish has always been there in science. And even back before there were professional scientists, when there were just wealthy amateurs.
Also, the great variety of early outlets like conferences, provide an opportunity to foster early disclosure of results because a scientist can disclose early and stake a claim, before having to get stuff through the peer review process.
Finally, for most scientists, the desire for publication outstrips the desire for patents. It's publications we use to measure our success; not patents.
I don't follow the argument. As you say, there's a market of scientific journals for a reason. There are too many articles for scientists to read. So we delegate the responsibility of pre-filtering to the journals. As you point out, we also delegate some of the responsibility of career-filtering to the journals. It's not a perfect system, but it works pretty well.
So why do we need to change our perception of scientific excellence before putting into place a peer-review mechanism? I don't follow the argument. Nor, for that matter, do I follow the bald assertion "This is very wrong, of course." Why? We need some measure to help us separate the wheat from the chaff? Why is the journal measure a particularly bad one?
Obviously, it can be gamed, but people are pretty savvy to obvious gaming tactics such as publishing in shoddy, uncritical journals.
Most normals don't want broadband because it's a nightmare from hell to bring up and keep up, at least based on my DSL experience. I would say I have had about 1 outage per 1.5 months on the average (over 2+ years). Each time I get to spend at least a half an hour on the phone with the fun folks at Qwest (sometimes much, much more). Plus you have to figure out how some obscure piece of routing hardware works.
This stuff is so far from Aunt-Tillie-ready it's not even funny.
Most normals would make a better tradeoff having dial-up that works instead of broadband that...uh, sorta works.... and requires a lot of care and feeding.
Actually, this is oversimplified and doesn't agree with US law. You have some very specific privacy rights, even when you're at work. For example, without some preparation and warning, your employer can't rifle your desk or listen to your phone calls. They can read your emails and look at anything on the computer they supply you.
The only reason for Frauenhofer to release the MP3 technology, instead of sitting on it, is because they think they can make money from it. They made the further decision that it would be better to release it free first, and then ambush us with charges later. OK, that's not very nice, but it still agrees with the purpose of patents. That is, it was the prospect of being able to ambush us with charges later that persuaded them to release the MP3 technique.
I don't like what they're doing either, but as I said it's not an argument against patent. Patent says "we'll give you a monopoly in exchange for disclosure." Frauenhofer has decided that the right way to use that monopoly is to get us all hooked up front, and then soak us later.
Just because they made the money after the release of the technology doesn't change the way patent works. It's the prospect of making money in the future that causes the disclosure.
Yes, we got sucker-punched. Yes, it's unethical. But no, this is not an argument against patents --- indeed, this case shows patents working. MP3s are all over the place precisely because Frauenhofer could be confident in their ability to sucker-punch us later, even though we all know how MP3s work.
Look at your own argument
If companies are just making money, rather than bringing new and good technologies to the market, then the purpose of the patent system is not being served. This is what Frauenhofer is doing.
This is not what Frauenhofer is doing. What they are (were) doing is (was) bringing us a new and good compression technique. The fact that they're still making money now does not contradict the patent concept. After all, even if you don't do the sucker-punch thing, you still get to make money for years after releasing your technology. A patent wouldn't be very useful if the monopoly expired at the time the patent was granted!
I've got to disagree with you there, although I hate to defend these patents.
Your argument has a clear logical hole in it --- in fact it shows that what Frauenhofer is doing is exactly what patents are for!
Obviously, the reason the inventors are embedding the ideas in standards (rather than holding them close to their chests) is precisely because they can wait until they are adopted and then cash in. That's exactly what the inventors of patents wanted! The alternative would be that the inventors would say "there's no prospect of us cashing in unless we hold these ideas to ourselves," they'd never get out into the world, except in closed devices, and we wouldn't all have cool mp3s.
You may think this is dirty pool, but it sure isn't an argument against the rationale for patents.
I didn't say anything about pointers. In fact, I'd favor leaving out that stuff for a while. Part of the reason I think that C-style languages are a poor choice for a first programming course.
In general, I'm a fan of introductory approaches that maintain abstraction (e.g., the SICP-style discussion in Scheme of how to do objects) and dislike things that introduce contingent facts about mapping onto hardware (e.g., procedural languages, where the first thing you have to learn is some rudimentary I/O, which is almost always a rat's nest of special cases, complexity and kludges).
For starters, I favor languages where you don't even see pointers!
I don't agree --- sequential, conditional and looping are the easy things. Teach them the hard stuff first, like recursion. The easy stuff can wait.
That's also the stuff that most of the foundation of computer science is built on, and which you need to know to be a good computer scientist.
I speak from hard experience, having tried to teach Theory of Computation to a class of college seniors who had never mastered recursion. Very ugly.
If you can get away with it, look at Abelson and Sussman's Structure and Interpretation of Computer Programs. Still the classic for concepts of programs and programming.
It depresses me to see how many of the comments on this article are complaints about how stupid the users are.
I think this really misses the point. If one or two people have trouble with strong passwords, ok, maybe. But if strong passwords are almost uniformly either not used, or written on post-its, then the solution is busted. It's not the job of tool-makers to demand that people change to fit the tools. We need to give users tools that they can use. If passwords suck, get something else.
I don't like it because it's got too many effing features! I was ok with the early versions, when it was just a register, some bill paying and the ability to reconcile. And it was rock solid. Now it has got a zillion features, most of whose existence I don't even know, and it fails every now and then.
I wish it was possible to get a UNIX-style program, that would do only one thing (manage a register), and do it well, and would then hook up easily with other programs, each of which would also do only one thing and do it well.
There was a story by Peter Maass in Slate some time ago that indicates this not to be just more internet blather. I quote:
Although the Israel-destroyed-the-World-Trade-Center theory is absurd, most people I speak to in Pakistan believe it. It's not just the headband rabble who sense the hand of the Mossad at work. I even heard the conspiracy theory during a visit to the posh Islamabad home of Ijzal ul-Haq, vice president of a major political party. As it happens, Ijzal ul-Haq is also the son of Pakistan's late military dictator Zia ul-Haq; an oil portrait of the old man hangs a few feet from the front door.
Also, though I am not able to find pointers (ok, I'm too lazy, fine), there were reports of a public opinion poll of the Islamic Middle East, that purported to find in no location, more that 40% willing to believe that Muslims were responsible for 9/11.
Then, of course, there are the Saudi state papers that publish reports about Jews making matzah out of the blood of Gentiles....
But where are internet consumers going to learn to develop these filters?
And what happens if those filters are simply "this is (is not) in agreement with [insert dogma here]?"
I'm sorry, but this doesn't look like much of a silver lining to me.
I remember reading in the Wall Street Journal an interview with a madrassah student who asserted proudly that dinosaurs were real, created by Jews and Americans to kill the faithful. How's that for a filter?
Invasive American culture -- from movies, music, fast-food -- have highlighted political and religious differences, from Europe to the Middle East and South Asia.
Is this claim really true? Seems like most of the anti-globalism comes out of Europe and the Middle East, and not so much out of South East Asia. Europe and South East Asia are heavily exposed to "invasive American culture." The Middle East? Not so much. There really is far less exposure to American culture in Saudi Arabia, home of lots of anti-Americanism, than there is in South East Asia, home of a lot less. So maybe this is just an argument that makes MacDonald's haters happy....
Ok, so I haven't eaten in MacDonald's for years, either. I still don't think Ronald created the Taliban.
If I was a large company converting to Linux, the initial step wouldn't be forking out an extra $50 or so for each machine to get the Codeweavers plugin. Your first step would be to shift office environments to one that is cross platform and moving all your file formats to that product.
Actually, that option is not available to many people. For example, if you are a government contractor, you have to worry about exchanging documents with your customers. Internal consistency on a non-MS platform is not an option, because internal consistency is not enough.
For those of us who prefer Linux, but need to read and write Office docs, this might be just the thing. And there would be a big savings over the status quo where I have to have two boxen.
This is a little bit of an overstatement. While the journals often are expensive, there are real costs besides copyediting, to running a paper journal that has to go through a printing press, etc.
Also, I think that this talk about reviewers stealing results and authors intentionally falsifying papers (pace Alan Sokal) is seriously oversold. I certainly haven't seen this problem in the refereed computer science journals I write and review for.
And while I like open publications, there are some things I miss about those old, expensive journals. Like the fact that they used to copy edit. Now I get journals filled with grammatical errors, if I'm lucky. If I'm not, they'll even have spelling errors. No one's checking them anymore.
I liked the intermediate phase of this revolution, when there still were edited peer-reviewed journals, but free publication was ensured by the availability of preprints and short, conference versions of papers.
I'm not a python user, but I use perl and Common Lisp a lot. I agree with this poster, why would you want to use C++, which is a low level language, to control high level languages? Wouldn't you be better off starting out with a high level language and then doing the high-cost bits in C/C++?
For that matter, the Kernighan and Plauger book (I think that's the one), although it reaches the opposite conclusion, clearly shows that you're probably just as well off writing perl as C++, for efficiency purposes.
In his follow-up article, Charles Connell correctly points out that software isn't harder to build than mechanical objects, it's easier.
This phenomenon was noted by Fred Brooks a long time ago in his book The Mythical Man-Month.
Interestingly, the fact that software is easy is part of the problem, not part of the solution:
The software entity is constantly subject to pressures for change. Of course, so are buildings, cars, computers. But manufactured things are infrequently changed after manufacture; they are superseded by later models, or essential changes are incorporated in later serial-number copies of the same basic design....Partly this is because the software in a system embodies its function, and the function is the part that most feels the pressure of change. Partly it is because software can be changed more easily --- it is pure thought-stuff, infinitely malleable. Buildings do in fact get changed, but the high costs of change, understood by all, serve to dampen the whims of the changers.
Brooks also points out that software usually bears the burden of conforming to the environment, again because it's easier to change the software than hardware aspects of the environment.
If you want to know why software stinks, Brooks' book, 20+ years later, is still the place to start.
I'm a pretty fanatic CL programmer and I agree that the lack of good GUI bindings is a real handicap.
IMHO, CLIM manages to be the worst of all worlds. It's crude, yet somehow bloated, too. And on top of all that, it's insanely expensive.
A good, bulletproof, free interface to Tk would get my vote.... Honestly, I wouldn't want CLIM even if it were free.
I've used Garnet and think it's slick as anything, but it's orphanware with no one carrying it on. Further, it requires mastering an entirely new, and different (and cool, I admit) object system. A significantly added cost on top of the already difficult sell of CL. Oh yeah, and Garnet out of the box is really, really, ugly. I had to sit down with a UI specialist for a long time to make my own version with OK-looking colors, roundtangles for buttons, etc., etc. Then it was way slick.
Why prefer XEmacs over GNU Emacs
on
GNU Emacs 21
·
· Score: 1
For me, as someone who maintains multiple emacsen across platforms, Xemacs has been handier because of the easier package management.
Is there any reason to revert to FSF emacs (aside from political)? I'd be willing to be put straight, but I can't see any reason to go back, so far.
I'm so glad that this has been posted to Slashdot! I'd been meaning to write a letter to Wired about this but was too lazy to get it done before the next issue came out....
The article has the Open Source movement all wrong. The author treats the Linux desktop issue as if it were IBM versus Microsoft, not Open Source versus Microsoft.The author spends most of the article lamenting that the Open Source movement is wasting its energies worrying about the desktop, when Microsoft owns it.
To put things in more pompous terms, the author spends a lot of time bemoaning the opportunity cost of spending time on the Linux desktop, and claiming that this time should all be spent on the server market.
This shows a total failure to understand the Open Source movement. The Open Source movement is not Open Source, Inc. Linus, Eric, RMS, or whoever is your pick for Open Source, Inc. CEO can't just say "yo! KDE-boy, toss in the towel on this desktop thing and spend more time on the server!" or "yo! I know you want to make a totally excellent system for tracking your MP3's, but you should be improving Apache, instead!"
People write Open Source software because they want to improve the tools that they use all the time. People who use desktops will want to improve the desktops and people who use servers will want to improve servers. And those people are not fungible --- they can't just be reassigned. And I think ESR's spot-on in the book when he talks about the fact that a lot of the quality of good Open Source software comes from the fact that people can't be reassigned to projects they don't give a rip about. People are passionate about the software they write, so they try to make it good.
The author's failure to understand this key fact about Open Source makes his whole perspective into nonsense.
The article might make sense as an argument about what Open Source packagers should be doing, but even there, the packagers are largely driven by software producers.
Those points are all well and good, but they don't get at my argument. I wasn't trying to argue for the abolition of C++. Just trying to point out some reasons for using Scheme over C++ for instructional purposes.
A CS degree says you can tackle big stuff and not get thrown the first day. If you don't want to do bread-and-butter data management work, but stuff more cutting edge, then yes, the compter as a scientific subject should be in your realm of knowledge. But when as many as 75% of any given CS program's students only want big-buck commercial data management gigs, why torture them with CS theory?
I've got an answer to this, from working with people like that. They don't know how to find a decent algorithm in a book. They reinvent everything and, pretty much universally, they aren't as smart as the people that invented them in the first place (that's not really fair --- they may be as smart, but without the benefit of knowledge created before them, they are helpless).
For example, I've seen people reinvent OO databases, I've seen more linear time searches than you can shake a stick at (this is a sure recipe for disaster in big-buck commercial data management), I've seen parsers that would make you want to cry, written by people who never took compilers, etc., etc.
CS without theory is like engineering without physics.
Slashdot Mirror
Yeah, but.... While I love Cryptonomicon it was written well past the time of Turing's life and seemed, at least to me, to have a relatively chirpy view of what it would like to be gay at that time. The Hodges biography is much better about that, although possibly inclined to view too much of Turing's life and work through that lens.
I've got to take issue with you on the claims you make. Sure, patents may have accelerated the desire to hold research closely, but the pressure to be the first to publish has always been there in science. And even back before there were professional scientists, when there were just wealthy amateurs.
Also, the great variety of early outlets like conferences, provide an opportunity to foster early disclosure of results because a scientist can disclose early and stake a claim, before having to get stuff through the peer review process.
Finally, for most scientists, the desire for publication outstrips the desire for patents. It's publications we use to measure our success; not patents.
I don't follow the argument. As you say, there's a market of scientific journals for a reason. There are too many articles for scientists to read. So we delegate the responsibility of pre-filtering to the journals. As you point out, we also delegate some of the responsibility of career-filtering to the journals. It's not a perfect system, but it works pretty well.
So why do we need to change our perception of scientific excellence before putting into place a peer-review mechanism? I don't follow the argument. Nor, for that matter, do I follow the bald assertion "This is very wrong, of course." Why? We need some measure to help us separate the wheat from the chaff? Why is the journal measure a particularly bad one?
Obviously, it can be gamed, but people are pretty savvy to obvious gaming tactics such as publishing in shoddy, uncritical journals.
Most normals don't want broadband because it's a nightmare from hell to bring up and keep up, at least based on my DSL experience. I would say I have had about 1 outage per 1.5 months on the average (over 2+ years). Each time I get to spend at least a half an hour on the phone with the fun folks at Qwest (sometimes much, much more). Plus you have to figure out how some obscure piece of routing hardware works.
This stuff is so far from Aunt-Tillie-ready it's not even funny.
Most normals would make a better tradeoff having dial-up that works instead of broadband that...uh, sorta works.... and requires a lot of care and feeding.
Actually, this is oversimplified and doesn't agree with US law. You have some very specific privacy rights, even when you're at work. For example, without some preparation and warning, your employer can't rifle your desk or listen to your phone calls. They can read your emails and look at anything on the computer they supply you.
I don't think you see my point.
The only reason for Frauenhofer to release the MP3 technology, instead of sitting on it, is because they think they can make money from it. They made the further decision that it would be better to release it free first, and then ambush us with charges later. OK, that's not very nice, but it still agrees with the purpose of patents. That is, it was the prospect of being able to ambush us with charges later that persuaded them to release the MP3 technique.
I don't like what they're doing either, but as I said it's not an argument against patent. Patent says "we'll give you a monopoly in exchange for disclosure." Frauenhofer has decided that the right way to use that monopoly is to get us all hooked up front, and then soak us later.
Just because they made the money after the release of the technology doesn't change the way patent works. It's the prospect of making money in the future that causes the disclosure.
Yes, we got sucker-punched. Yes, it's unethical. But no, this is not an argument against patents --- indeed, this case shows patents working. MP3s are all over the place precisely because Frauenhofer could be confident in their ability to sucker-punch us later, even though we all know how MP3s work.
Look at your own argument
This is not what Frauenhofer is doing. What they are (were) doing is (was) bringing us a new and good compression technique. The fact that they're still making money now does not contradict the patent concept. After all, even if you don't do the sucker-punch thing, you still get to make money for years after releasing your technology. A patent wouldn't be very useful if the monopoly expired at the time the patent was granted!
I've got to disagree with you there, although I hate to defend these patents.
Your argument has a clear logical hole in it --- in fact it shows that what Frauenhofer is doing is exactly what patents are for!
Obviously, the reason the inventors are embedding the ideas in standards (rather than holding them close to their chests) is precisely because they can wait until they are adopted and then cash in. That's exactly what the inventors of patents wanted! The alternative would be that the inventors would say "there's no prospect of us cashing in unless we hold these ideas to ourselves," they'd never get out into the world, except in closed devices, and we wouldn't all have cool mp3s.
You may think this is dirty pool, but it sure isn't an argument against the rationale for patents.
Huh? Pointers?
I didn't say anything about pointers. In fact, I'd favor leaving out that stuff for a while. Part of the reason I think that C-style languages are a poor choice for a first programming course.
In general, I'm a fan of introductory approaches that maintain abstraction (e.g., the SICP-style discussion in Scheme of how to do objects) and dislike things that introduce contingent facts about mapping onto hardware (e.g., procedural languages, where the first thing you have to learn is some rudimentary I/O, which is almost always a rat's nest of special cases, complexity and kludges).
For starters, I favor languages where you don't even see pointers!
I don't agree --- sequential, conditional and looping are the easy things. Teach them the hard stuff first, like recursion. The easy stuff can wait.
That's also the stuff that most of the foundation of computer science is built on, and which you need to know to be a good computer scientist.
I speak from hard experience, having tried to teach Theory of Computation to a class of college seniors who had never mastered recursion. Very ugly.
If you can get away with it, look at Abelson and Sussman's Structure and Interpretation of Computer Programs. Still the classic for concepts of programs and programming.
It depresses me to see how many of the comments on this article are complaints about how stupid the users are.
I think this really misses the point. If one or two people have trouble with strong passwords, ok, maybe. But if strong passwords are almost uniformly either not used, or written on post-its, then the solution is busted. It's not the job of tool-makers to demand that people change to fit the tools. We need to give users tools that they can use. If passwords suck, get something else.
I don't like it because it's got too many effing features! I was ok with the early versions, when it was just a register, some bill paying and the ability to reconcile. And it was rock solid. Now it has got a zillion features, most of whose existence I don't even know, and it fails every now and then.
I wish it was possible to get a UNIX-style program, that would do only one thing (manage a register), and do it well, and would then hook up easily with other programs, each of which would also do only one thing and do it well.
Grumble. I hate bloatware.
Actually, we're both right.
I believe that you are right, that the article from Saudi Arabia claimed that Hamentashen are made with blood.
But the "standard form" blood libel is about matzah, not purim cakes.
this kind of literary criticism of hate literature is pretty weird....
There was a story by Peter Maass in Slate some time ago that indicates this not to be just more internet blather. I quote:
Also, though I am not able to find pointers (ok, I'm too lazy, fine), there were reports of a public opinion poll of the Islamic Middle East, that purported to find in no location, more that 40% willing to believe that Muslims were responsible for 9/11.
Then, of course, there are the Saudi state papers that publish reports about Jews making matzah out of the blood of Gentiles....
But where are internet consumers going to learn to develop these filters?
And what happens if those filters are simply "this is (is not) in agreement with [insert dogma here]?"
I'm sorry, but this doesn't look like much of a silver lining to me.
I remember reading in the Wall Street Journal an interview with a madrassah student who asserted proudly that dinosaurs were real, created by Jews and Americans to kill the faithful. How's that for a filter?
Is this claim really true? Seems like most of the anti-globalism comes out of Europe and the Middle East, and not so much out of South East Asia. Europe and South East Asia are heavily exposed to "invasive American culture." The Middle East? Not so much. There really is far less exposure to American culture in Saudi Arabia, home of lots of anti-Americanism, than there is in South East Asia, home of a lot less. So maybe this is just an argument that makes MacDonald's haters happy....
Ok, so I haven't eaten in MacDonald's for years, either. I still don't think Ronald created the Taliban.
Actually, that option is not available to many people. For example, if you are a government contractor, you have to worry about exchanging documents with your customers. Internal consistency on a non-MS platform is not an option, because internal consistency is not enough.
For those of us who prefer Linux, but need to read and write Office docs, this might be just the thing. And there would be a big savings over the status quo where I have to have two boxen.
This is a little bit of an overstatement. While the journals often are expensive, there are real costs besides copyediting, to running a paper journal that has to go through a printing press, etc.
Also, I think that this talk about reviewers stealing results and authors intentionally falsifying papers (pace Alan Sokal) is seriously oversold. I certainly haven't seen this problem in the refereed computer science journals I write and review for.
And while I like open publications, there are some things I miss about those old, expensive journals. Like the fact that they used to copy edit. Now I get journals filled with grammatical errors, if I'm lucky. If I'm not, they'll even have spelling errors. No one's checking them anymore.
I liked the intermediate phase of this revolution, when there still were edited peer-reviewed journals, but free publication was ensured by the availability of preprints and short, conference versions of papers.
For people who don't want to read an entire book on the subject, Powers wrote a very interesting review of the play Copenhagen in the New York Review.
I'm not a python user, but I use perl and Common Lisp a lot. I agree with this poster, why would you want to use C++, which is a low level language, to control high level languages? Wouldn't you be better off starting out with a high level language and then doing the high-cost bits in C/C++?
For that matter, the Kernighan and Plauger book (I think that's the one), although it reaches the opposite conclusion, clearly shows that you're probably just as well off writing perl as C++, for efficiency purposes.
In his follow-up article, Charles Connell correctly points out that software isn't harder to build than mechanical objects, it's easier.
This phenomenon was noted by Fred Brooks a long time ago in his book The Mythical Man-Month.
Interestingly, the fact that software is easy is part of the problem, not part of the solution:
Brooks also points out that software usually bears the burden of conforming to the environment, again because it's easier to change the software than hardware aspects of the environment.
If you want to know why software stinks, Brooks' book, 20+ years later, is still the place to start.
I'm a pretty fanatic CL programmer and I agree that the lack of good GUI bindings is a real handicap.
IMHO, CLIM manages to be the worst of all worlds. It's crude, yet somehow bloated, too. And on top of all that, it's insanely expensive.
A good, bulletproof, free interface to Tk would get my vote.... Honestly, I wouldn't want CLIM even if it were free.
I've used Garnet and think it's slick as anything, but it's orphanware with no one carrying it on. Further, it requires mastering an entirely new, and different (and cool, I admit) object system. A significantly added cost on top of the already difficult sell of CL. Oh yeah, and Garnet out of the box is really, really, ugly. I had to sit down with a UI specialist for a long time to make my own version with OK-looking colors, roundtangles for buttons, etc., etc. Then it was way slick.
For me, as someone who maintains multiple emacsen across platforms, Xemacs has been handier because of the easier package management.
Is there any reason to revert to FSF emacs (aside from political)? I'd be willing to be put straight, but I can't see any reason to go back, so far.
I'm so glad that this has been posted to Slashdot! I'd been meaning to write a letter to Wired about this but was too lazy to get it done before the next issue came out....
The article has the Open Source movement all wrong. The author treats the Linux desktop issue as if it were IBM versus Microsoft, not Open Source versus Microsoft.The author spends most of the article lamenting that the Open Source movement is wasting its energies worrying about the desktop, when Microsoft owns it.
To put things in more pompous terms, the author spends a lot of time bemoaning the opportunity cost of spending time on the Linux desktop, and claiming that this time should all be spent on the server market.
This shows a total failure to understand the Open Source movement. The Open Source movement is not Open Source, Inc. Linus, Eric, RMS, or whoever is your pick for Open Source, Inc. CEO can't just say "yo! KDE-boy, toss in the towel on this desktop thing and spend more time on the server!" or "yo! I know you want to make a totally excellent system for tracking your MP3's, but you should be improving Apache, instead!"
People write Open Source software because they want to improve the tools that they use all the time. People who use desktops will want to improve the desktops and people who use servers will want to improve servers. And those people are not fungible --- they can't just be reassigned. And I think ESR's spot-on in the book when he talks about the fact that a lot of the quality of good Open Source software comes from the fact that people can't be reassigned to projects they don't give a rip about. People are passionate about the software they write, so they try to make it good.
The author's failure to understand this key fact about Open Source makes his whole perspective into nonsense.
The article might make sense as an argument about what Open Source packagers should be doing, but even there, the packagers are largely driven by software producers.
Those points are all well and good, but they don't get at my argument. I wasn't trying to argue for the abolition of C++. Just trying to point out some reasons for using Scheme over C++ for instructional purposes.
For example, I've seen people reinvent OO databases, I've seen more linear time searches than you can shake a stick at (this is a sure recipe for disaster in big-buck commercial data management), I've seen parsers that would make you want to cry, written by people who never took compilers, etc., etc.
CS without theory is like engineering without physics.