Basically, the Art of Computer Programming. I can't read MIX, so the source code listings have no meaning to me. I agree that if we are talking about actual programs, the source code is a prime source for information on a program (although I doubt that this a good thing), but the fundamental ideas behind a program are not expressed in a programming language, but usually in more-or-less plain English.
Or look at the typical documentation of a library interface (if there is any). Usually, the interface itself is described in terms of the programming language, but the semantics are described in English. Of course, the documentation might include source code examples, but usually, these show how to use the interface, and not what the library does.
(If you want to read it in a programming language,
you can check the library source code after all.)
If the copyright is not exercised, an author can regain it from the publisher, I think. (Mind that the source code, as a form of expression, is not published.)
In addition, the Berne Convention has provisions for the mutilation of works of art (and other moral rights). If programmers start to use this against their employers, we're going to face interesting times.
[C]omputer source code, though unintelligible to many, is the preferred method of communication among computer programmers. Because computer source code is an espressive means for the exchagne of information and ideas about computer programming, we hold that it it protected by the first amendment.
And proprietary software is censorship of the employed programmers? Oh dear.
Computer programs are intended primarly for human-machine communication, at least that's the way most programming languages are designed. Programmers communicate among each other using documentation (at least in theory). Of course, in almost all cases, there isn't much usable, up-to-date documentation, and this leads to poor software quality: the next programmer, due to lack of documentation, does not understand the design decisions of his predecessor, and thus can't walk in his footsteps.
And, regarding the second part, I very much doubt that an article describing a new idea about computer programming would pass the review process if it was written in, say, C. Hardly anybody expresses abstract ideas in concrete programming languages.
If we were all using literary programming tools such as CWEB, the observation would have some basis, but currently, such tools are used only infrequently. Perhaps we should change that, in order to gain a bit of constitutional protection. (Unfortunately, here in Germany, Free Speech is not protected, you only have the right to utter your opinion, and this right is severely limited by law in comparision to the US situation.)
Banning cookies might get unexpected support
on
EU May Outlaw Cookies
·
· Score: 3, Informative
Banning cookies might get unexpected support: from the law enforcement camp. After all, if cookies are no longer permitted, those interesting session IDs have to be placed in the requested URIs. And these URIs are logged all over the place: by the web server itself, by proxies along the way, by the browser (in theory, session cookies should expire when then browser is terminated). So banning cookies makes session tracing much easier for everyone but the actual web server developer.
Cookies, when used in a responsible way, can increase privacy. Of course, that is not true with those practically eternal cookies which expire some day in the year 2037 or so. On the other hand, there are other tracing methods such as exclusively dynamic URIs or even cache timing attacks (yet another interesting Felten paper, BTW).
In my opinion, you should not outlaw the tool, but the intention to gather data. Recently, we've seen so many attempts at restricting tools which have some negative potential, competely neglecting the positive possibilities such tools present. Shall we make the same mistake again?
C is defined for a binary machine, like most programming languages (if they are defined properly at all). That's why it's hard to believe that we're going to see non-binary computing anytime soon. It would be very inefficient to reuse old software, making the theoretical effiency gain rather worthless. (Hmm, but this didn't prevent Intel/HP/et al. from developing the IA-64 architecture, may be they start trinary computing next?).
By the way, Ada does have support for trit operations (in some bizarre way), but this was merely an accident.
There are still reports that reiserfsck doesn't work as desired, only as specified. (It succeeds in putting the file system in a well-defined state, but alas, sometimes, all files are gone.) However, ReiserFS is in the official tree, so it gets most testing, and it directly follows other kernel changes, so the kernel component should work fairly reliably.
XFS seems to have more mature userland tools (file system check, dump/restore). However, the XFS developers have a hard time tracking the mainline kernel, so if you want to get all those security fixes, you have to use their CVS version which gets less testing and is by no means an official version.
If the file system checks are in the 10 minute range, this is no problem, but if he's serving a couple of hundred gigabytes, this wouldn't be the case.
In any case, if the file system is not that large and you want a solution that simply works, use ext2. All journaling file systems are not stable yet or lack essential tools (such as dump/restore or working fsck for real emergencies) or both. Of course, running some advanced file system is interesting in itself, so you might want to try one of them even if you don't need it, strictly speaking.
Microsoft has done this before. Here in Germany, we've got a Bavarian politician with the name of "Gauweiler" (he's as much right wing as you can while being a member of a democratic party, and sometimes he even agitates against parts of the German constitution). In one of the early Windows versions, the spellchecker suggested a correction, "Gauleiter", for "Gauweiler". Since some regional Nazi adminstrators were called "Gauleiter" during the Third Reich, he was not amused, and he and his party protested. In the end, Microsoft (or the company from which Microsoft licensed the spellchecker) removed the word "Gauleiter" from the dictionary, thus eliminating the problem.
Maybe such experiences motivated Microsoft to remove some naughty words from their English (or do you mean American?;-) dictionary.
BTW, at the beginning of the nineties, people still thought that spellcheckers were more funny than useful. For example, the German version suggested "Spanner" (peeping tom) for "Scanner" (a digital imaging device which can also be used digitize porn).
As far as I can tell, 2.2.20pre10 is still vulnerable to the problems with ptrace() and symlinks (the patches in the advisory have not been applied).
Shameless plug:
There are also German summaries for the ptrace() and symlinks. Well, with the recent advisory on the security problem in Windows Media Player (regarding DRM), I shouldn't travel to the US while the DMCA is in place.
The server I use to access my bank account is running Apache. When the firewall settings were less tight, you could see that they were running a Linux kernel somewhere (at least for the firewalling), now I'm not sure.
To be honest, the bank is a "Genossenschaft", a concept which was invented back in the 19th century and is in some way similar to the free software movement. The bank itself uses OS/2 almost exclusively on desktops, which means that they were unaffected by the recent Microsoft worm craze, and they don't have a direct Internet connection. AFAIK, the computing center still issues IP addresses according to a scheme which is not compatible with the official IANA one, i.e. they use the whole IPv4 address space as if there was no public Internet.
I think he gives all scenarios a credibility rating of 10/10, even the Independence Day scenario. This guy must live in a different world. However, I've bought his predictions book, and I plan to read in twenty years time or so. It will certainly be funny.
Why? The Linux kernel had a similar problem with ptrace() which was fixed at the beginning of this year, and Windows didn't become more secure all of a sudden.
And after all, hardly anyone at the Windows camp is concerned about local security. Quite a lot of hardware and software runs only under an administrative account and needs write access to the system directories (and this software is still called "Windows compiatble"). So, from a practical point of view, you are often forced to run your Windows machine with very liberal security policies, not comparable to the standard UNIX setup. And if you don't you have got a long way to go from the default Microsoft setup, which focuses far too much on compatibility with former Windows versions.
And, by the way, it is not reasonable to expect that general purpose UNIX-like operating systems can withstand attacks from local users. Obviously, these systems were not designed with this kind of threat in mind.
That's partly because Macs used to have a basically non-preemptive operating system. Usually, this is a severe limitation, but in some cases, you can get lower latency and more exact timing. On the consumer Windows platform, too, there were some reports that MIDI performance degraded because of partly preemptive between Win32 applications introduced by Windows 95.
However, with MacOS X, you've got a real preemptive scheduler, so this advantage is probably lost forever. There may be some real-time scheduling options, but scheduling a typical UNIX workload and real-time scheduling don't mix very well (although there are some hackish solutions).
Why has this to be a good thing for the German economy? There may be many small software manufacturers over here which produce all kinds of stuff, but hardly any larger ones (except SAP, of course). Star Division got bought out by Sun, and I don't know if SoftMaker is still independent (they offer some kind of office suite, too, but it's proprietary software). This means it's difficult to find someone who can actually deliver the software you need (and whose company is not owned by a US-based multinational).
Now, if you turn to the free software companies: there aren't many of them either. Very few a focused on free software, but most (like SuSE) take free software and bundle it with proprietary enhancments.
Of course, chosing free software for the Bundestag is a political message, and I believe that free software benefits everyone in the long term (like pollution control, which has got a short to middle term detrimental effect, too). But at the moment, if the Bundestag would like to do something for the economy, not chosing free software would probably effect software companies more positively and noticeably (as long you actually keep most contracts to German companies!). I don't think many software companies over here are ready to enter the adventure world of free software -- yet.
In some states, they are. For example, in Virginia, shrink-wrapped licenses can result in legally binding contracts, at least I think so.
However, in other countries (such as Germany), the EULAs are invalid twice: first, as an end user, you don't license the software over here (and you don't need a license, copyright law gives you the right to run your own copy of the software), you simply buy it at a shop (so there's no contractual bond between you and Microsoft), and second, we have quite strong consumer protection laws, according to which most provisions of the EULA are null and void (even if were part of the contract between you and the software dealer, which they aren't).
Microsoft even claimed that installing an upgrade makes it illegal to use the previous version (even on the same computer, by the same person). Why do they call it "upgrade", if they really mean "replacement"?
Fortunately, if you buy some COTS software here in Germany, you don't get a license to use the software, you only enter into contract with the software dealer (who is not allowed to license software, only to sell copies). This means that you have no right to use the software in any way, except for the default rights granted to you by copyright law (we've got a rather strange construction over here, it has to do with the fact that running a computer program involves making at least one copy of the program). And these rights have nothing to do with "upgrades". I think you are even permitted to sell the original version if you destroy all copies of it, and use the upgrade afterwards (if it still runs without these copies).
BTW, back in 1995 or so, Microsoft allowed you to use the same license for Office at your workplace and at home (but this permission was restricted to application software, operating systems were not included).
You first have to examine carfully the chracter set your current application can deal with. Is it ASCII? Or just the printable range? Or do most routines treat everything as sequences of 8-bit characters? Is the null character permitted in data? And so on.
After that, you have to identify the operations which are character set specific. This can be quite a bit of work. Character set specific operations include case conversion, collating, normalizing, measuring string length and character width (for formatting plain text), text rendering in general, and so on.
Now you look at your tools. Do they prefer some kind of Unicode encoding? For example, with Java or Windows, using UTF-16 is most convinient (some would say: mandated).
Now you put the pieces together and look for a suitable internal representation (not necessarily "Unicode", i.e. UTF-8, UTF-16, or UTF-32), identify points at which data has to be converted (usually, it is a good idea to minimize this, but if you want to fit everything together, there is sometimes no other choice), and modules and external tools which have to be replaced because adjusting them or adapting to them is too much work.
Your web page generation tools probably need a complete overhaul, so that they are able to minimize the charset being used (for example, German text is sent as ISO-8859-1, but Russian text as KOI8-R or something like that), since client-side Unicode support is mostly ready, but many people don't have the necessary fonts.
I think the GnuPG contract has expired now, and funding was rather symbolic (approximately USD 100,000). However, there's a new free software project related to end-user cryptography standards which is supported by the federal government. There is clearly some interest in free software, but I've got the impression that publicly funded software development (be it free software or not) is restricted to areas in which is strictly necessary (for example, databases for use by police forces), otherwise COTS software is used, usually by Microsoft. We don't have government agencies which have lots of money to spend on projects which result in technology which can be used by the general public later on. The German Ministry of Defense has a rather small research budget, too (if they have any at all). OTOH, our federal parliament (the Bundestag) is investigating a migration to GNU/Linux on desktops, but this falls more or less into the COTS category, too, I think.
So, when viewed with this additional information, even limited public funding for free software is probably significant for the government. In addition, some government official offered strong moral support for free software, and there is even a brochure on free software targeted at SMEs. However, other government officials don't care a bit about free software and push Microsoft contracts and software patents alike. Things are not very consistent over here, and it's still not clear which side will dominate in a few years.
Yes, for example, in the past, the DoD provided initial funding for the development of a GCC front end. Of course, you have already guessed, for which language: Ada. These days, this front end (called GNAT) has finally become part of the official GCC sources, BTW.
The DoD even mandated that the source code is licensed under the GPL. The front end is reasonably independent of the GCC backend (JGNAT which targets the Java virtual machine exists, and it doesn't use the GCC backend), so this makes indeed a difference even though GCC itself is GPLed. (The developers, however, would have chosen to use the GPL even without being forced to do so.)
I'm sorry to say that, but the German federal government (and some governments in the Länder) tends to enter into long-term contracts with proprietary software vendors, especially with Microsoft.
And BSI does not always support open standards, especially in key cryptography areas. BSI tried to convince us to use Chiasmus for Windows, a closed-source implementation of their proprietary Chiasmus cipher. AFAIK, a request to hand over information on which we could base a decision was never answered.
I would have expected the the article dealt with the strong tendency in the Microsoft Windows camp to ignore MIME content types in favor of file extensions. You have to keep this in mind constantly, especially since there is still some software (admittedly, not by Microsoft) which truncates a file name to 11 characters and uses the last three characters as extension to decide which application to start. If you don't pay attention, you can easily compose a message which people running Microsoft Windows can't read. Perhaps this is evil conspiracy against all these *NIX people, to make them to appear as if they can't properly send email with attachments?
And the claim that the MacOS is superior in this context is a bit awkward: it's much more difficult to change the File Type/Creator fields with a Mac. I even doubt there is a central registry in which you can make an entry that all your Adobe Illustrator shall be opened from now on by Corel Draw (yikes).
A huge change with MMIX support has very recently been committed to the binutils CVS, and a MMIX is about to be integrated in the GCC CVS. Yeah, Ada on MMIX, an interesting combination.
Some time ago, I bought an "IBM-DTTA-371440" (sorry, don't know the trivial name). It was advertised as desktop hard disk (by the DeskStar label), but according to the manual, it requires a constant air flow on the top, and I have practical experience that this is true.
I think you can't run most 7200 RPM drives inside a desktop chassis, at least not if you've got several other heat sources in it.
Slashdot Mirror
Or look at the typical documentation of a library interface (if there is any). Usually, the interface itself is described in terms of the programming language, but the semantics are described in English. Of course, the documentation might include source code examples, but usually, these show how to use the interface, and not what the library does. (If you want to read it in a programming language, you can check the library source code after all.)
If the copyright is not exercised, an author can regain it from the publisher, I think. (Mind that the source code, as a form of expression, is not published.)
In addition, the Berne Convention has provisions for the mutilation of works of art (and other moral rights). If programmers start to use this against their employers, we're going to face interesting times.
Computer programs are intended primarly for human-machine communication, at least that's the way most programming languages are designed. Programmers communicate among each other using documentation (at least in theory). Of course, in almost all cases, there isn't much usable, up-to-date documentation, and this leads to poor software quality: the next programmer, due to lack of documentation, does not understand the design decisions of his predecessor, and thus can't walk in his footsteps.
And, regarding the second part, I very much doubt that an article describing a new idea about computer programming would pass the review process if it was written in, say, C. Hardly anybody expresses abstract ideas in concrete programming languages.
If we were all using literary programming tools such as CWEB, the observation would have some basis, but currently, such tools are used only infrequently. Perhaps we should change that, in order to gain a bit of constitutional protection. (Unfortunately, here in Germany, Free Speech is not protected, you only have the right to utter your opinion, and this right is severely limited by law in comparision to the US situation.)
Cookies, when used in a responsible way, can increase privacy. Of course, that is not true with those practically eternal cookies which expire some day in the year 2037 or so. On the other hand, there are other tracing methods such as exclusively dynamic URIs or even cache timing attacks (yet another interesting Felten paper, BTW).
In my opinion, you should not outlaw the tool, but the intention to gather data. Recently, we've seen so many attempts at restricting tools which have some negative potential, competely neglecting the positive possibilities such tools present. Shall we make the same mistake again?
C is defined for a binary machine, like most programming languages (if they are defined properly at all). That's why it's hard to believe that we're going to see non-binary computing anytime soon. It would be very inefficient to reuse old software, making the theoretical effiency gain rather worthless. (Hmm, but this didn't prevent Intel/HP/et al. from developing the IA-64 architecture, may be they start trinary computing next?).
By the way, Ada does have support for trit operations (in some bizarre way), but this was merely an accident.
There are still reports that reiserfsck doesn't work as desired, only as specified. (It succeeds in putting the file system in a well-defined state, but alas, sometimes, all files are gone.) However, ReiserFS is in the official tree, so it gets most testing, and it directly follows other kernel changes, so the kernel component should work fairly reliably.
XFS seems to have more mature userland tools (file system check, dump/restore). However, the XFS developers have a hard time tracking the mainline kernel, so if you want to get all those security fixes, you have to use their CVS version which gets less testing and is by no means an official version.
If the file system checks are in the 10 minute range, this is no problem, but if he's serving a couple of hundred gigabytes, this wouldn't be the case.
In any case, if the file system is not that large and you want a solution that simply works, use ext2. All journaling file systems are not stable yet or lack essential tools (such as dump/restore or working fsck for real emergencies) or both. Of course, running some advanced file system is interesting in itself, so you might want to try one of them even if you don't need it, strictly speaking.
Microsoft has done this before. Here in Germany, we've got a Bavarian politician with the name of "Gauweiler" (he's as much right wing as you can while being a member of a democratic party, and sometimes he even agitates against parts of the German constitution). In one of the early Windows versions, the spellchecker suggested a correction, "Gauleiter", for "Gauweiler". Since some regional Nazi adminstrators were called "Gauleiter" during the Third Reich, he was not amused, and he and his party protested. In the end, Microsoft (or the company from which Microsoft licensed the spellchecker) removed the word "Gauleiter" from the dictionary, thus eliminating the problem.
;-) dictionary.
Maybe such experiences motivated Microsoft to remove some naughty words from their English (or do you mean American?
BTW, at the beginning of the nineties, people still thought that spellcheckers were more funny than useful. For example, the German version suggested "Spanner" (peeping tom) for "Scanner" (a digital imaging device which can also be used digitize porn).
Shameless plug: There are also German summaries for the ptrace() and symlinks. Well, with the recent advisory on the security problem in Windows Media Player (regarding DRM), I shouldn't travel to the US while the DMCA is in place.
The server I use to access my bank account is running Apache. When the firewall settings were less tight, you could see that they were running a Linux kernel somewhere (at least for the firewalling), now I'm not sure.
To be honest, the bank is a "Genossenschaft", a concept which was invented back in the 19th century and is in some way similar to the free software movement. The bank itself uses OS/2 almost exclusively on desktops, which means that they were unaffected by the recent Microsoft worm craze, and they don't have a direct Internet connection. AFAIK, the computing center still issues IP addresses according to a scheme which is not compatible with the official IANA one, i.e. they use the whole IPv4 address space as if there was no public Internet.
I think he gives all scenarios a credibility rating of 10/10, even the Independence Day scenario. This guy must live in a different world. However, I've bought his predictions book, and I plan to read in twenty years time or so. It will certainly be funny.
Why? The Linux kernel had a similar problem with ptrace() which was fixed at the beginning of this year, and Windows didn't become more secure all of a sudden.
And after all, hardly anyone at the Windows camp is concerned about local security. Quite a lot of hardware and software runs only under an administrative account and needs write access to the system directories (and this software is still called "Windows compiatble"). So, from a practical point of view, you are often forced to run your Windows machine with very liberal security policies, not comparable to the standard UNIX setup. And if you don't you have got a long way to go from the default Microsoft setup, which focuses far too much on compatibility with former Windows versions.
And, by the way, it is not reasonable to expect that general purpose UNIX-like operating systems can withstand attacks from local users. Obviously, these systems were not designed with this kind of threat in mind.
That's partly because Macs used to have a basically non-preemptive operating system. Usually, this is a severe limitation, but in some cases, you can get lower latency and more exact timing. On the consumer Windows platform, too, there were some reports that MIDI performance degraded because of partly preemptive between Win32 applications introduced by Windows 95.
However, with MacOS X, you've got a real preemptive scheduler, so this advantage is probably lost forever. There may be some real-time scheduling options, but scheduling a typical UNIX workload and real-time scheduling don't mix very well (although there are some hackish solutions).
Why has this to be a good thing for the German economy? There may be many small software manufacturers over here which produce all kinds of stuff, but hardly any larger ones (except SAP, of course). Star Division got bought out by Sun, and I don't know if SoftMaker is still independent (they offer some kind of office suite, too, but it's proprietary software). This means it's difficult to find someone who can actually deliver the software you need (and whose company is not owned by a US-based multinational).
Now, if you turn to the free software companies: there aren't many of them either. Very few a focused on free software, but most (like SuSE) take free software and bundle it with proprietary enhancments.
Of course, chosing free software for the Bundestag is a political message, and I believe that free software benefits everyone in the long term (like pollution control, which has got a short to middle term detrimental effect, too). But at the moment, if the Bundestag would like to do something for the economy, not chosing free software would probably effect software companies more positively and noticeably (as long you actually keep most contracts to German companies!). I don't think many software companies over here are ready to enter the adventure world of free software -- yet.
In some states, they are. For example, in Virginia, shrink-wrapped licenses can result in legally binding contracts, at least I think so.
However, in other countries (such as Germany), the EULAs are invalid twice: first, as an end user, you don't license the software over here (and you don't need a license, copyright law gives you the right to run your own copy of the software), you simply buy it at a shop (so there's no contractual bond between you and Microsoft), and second, we have quite strong consumer protection laws, according to which most provisions of the EULA are null and void (even if were part of the contract between you and the software dealer, which they aren't).
Microsoft even claimed that installing an upgrade makes it illegal to use the previous version (even on the same computer, by the same person). Why do they call it "upgrade", if they really mean "replacement"?
Fortunately, if you buy some COTS software here in Germany, you don't get a license to use the software, you only enter into contract with the software dealer (who is not allowed to license software, only to sell copies). This means that you have no right to use the software in any way, except for the default rights granted to you by copyright law (we've got a rather strange construction over here, it has to do with the fact that running a computer program involves making at least one copy of the program). And these rights have nothing to do with "upgrades". I think you are even permitted to sell the original version if you destroy all copies of it, and use the upgrade afterwards (if it still runs without these copies).
BTW, back in 1995 or so, Microsoft allowed you to use the same license for Office at your workplace and at home (but this permission was restricted to application software, operating systems were not included).
You first have to examine carfully the chracter set your current application can deal with. Is it ASCII? Or just the printable range? Or do most routines treat everything as sequences of 8-bit characters? Is the null character permitted in data? And so on.
After that, you have to identify the operations which are character set specific. This can be quite a bit of work. Character set specific operations include case conversion, collating, normalizing, measuring string length and character width (for formatting plain text), text rendering in general, and so on.
Now you look at your tools. Do they prefer some kind of Unicode encoding? For example, with Java or Windows, using UTF-16 is most convinient (some would say: mandated).
Now you put the pieces together and look for a suitable internal representation (not necessarily "Unicode", i.e. UTF-8, UTF-16, or UTF-32), identify points at which data has to be converted (usually, it is a good idea to minimize this, but if you want to fit everything together, there is sometimes no other choice), and modules and external tools which have to be replaced because adjusting them or adapting to them is too much work.
Your web page generation tools probably need a complete overhaul, so that they are able to minimize the charset being used (for example, German text is sent as ISO-8859-1, but Russian text as KOI8-R or something like that), since client-side Unicode support is mostly ready, but many people don't have the necessary fonts.
I think the GnuPG contract has expired now, and funding was rather symbolic (approximately USD 100,000). However, there's a new free software project related to end-user cryptography standards which is supported by the federal government. There is clearly some interest in free software, but I've got the impression that publicly funded software development (be it free software or not) is restricted to areas in which is strictly necessary (for example, databases for use by police forces), otherwise COTS software is used, usually by Microsoft. We don't have government agencies which have lots of money to spend on projects which result in technology which can be used by the general public later on. The German Ministry of Defense has a rather small research budget, too (if they have any at all). OTOH, our federal parliament (the Bundestag) is investigating a migration to GNU/Linux on desktops, but this falls more or less into the COTS category, too, I think.
So, when viewed with this additional information, even limited public funding for free software is probably significant for the government. In addition, some government official offered strong moral support for free software, and there is even a brochure on free software targeted at SMEs. However, other government officials don't care a bit about free software and push Microsoft contracts and software patents alike. Things are not very consistent over here, and it's still not clear which side will dominate in a few years.
Yes, for example, in the past, the DoD provided initial funding for the development of a GCC front end. Of course, you have already guessed, for which language: Ada. These days, this front end (called GNAT) has finally become part of the official GCC sources, BTW.
The DoD even mandated that the source code is licensed under the GPL. The front end is reasonably independent of the GCC backend (JGNAT which targets the Java virtual machine exists, and it doesn't use the GCC backend), so this makes indeed a difference even though GCC itself is GPLed. (The developers, however, would have chosen to use the GPL even without being forced to do so.)
I'm sorry to say that, but the German federal government (and some governments in the Länder) tends to enter into long-term contracts with proprietary software vendors, especially with Microsoft.
And BSI does not always support open standards, especially in key cryptography areas. BSI tried to convince us to use Chiasmus for Windows, a closed-source implementation of their proprietary Chiasmus cipher. AFAIK, a request to hand over information on which we could base a decision was never answered.
I would have expected the the article dealt with the strong tendency in the Microsoft Windows camp to ignore MIME content types in favor of file extensions. You have to keep this in mind constantly, especially since there is still some software (admittedly, not by Microsoft) which truncates a file name to 11 characters and uses the last three characters as extension to decide which application to start. If you don't pay attention, you can easily compose a message which people running Microsoft Windows can't read. Perhaps this is evil conspiracy against all these *NIX people, to make them to appear as if they can't properly send email with attachments?
And the claim that the MacOS is superior in this context is a bit awkward: it's much more difficult to change the File Type/Creator fields with a Mac. I even doubt there is a central registry in which you can make an entry that all your Adobe Illustrator shall be opened from now on by Corel Draw (yikes).
There's a simulator in Knuth's book. (It's source code is available for download, but you need CWEB, of course.)
A huge change with MMIX support has very recently been committed to the binutils CVS, and a MMIX is about to be integrated in the GCC CVS. Yeah, Ada on MMIX, an interesting combination.
MMIX, of course.
Some time ago, I bought an "IBM-DTTA-371440" (sorry, don't know the trivial name). It was advertised as desktop hard disk (by the DeskStar label), but according to the manual, it requires a constant air flow on the top, and I have practical experience that this is true.
I think you can't run most 7200 RPM drives inside a desktop chassis, at least not if you've got several other heat sources in it.