Not at all. NT has been available since 1993, and Windows 3.x/9x deliberatly sacrifice robustness for compatibility and flexibility (since thats what the vast majority of home users want). Thats why NT/2000 has long been the recommended desktop for business users (where stability is presumably more important than the ability to use old games, etc.), while 9x was designed primarily for home users (where users would rather have the system crash from time to time than lose their old software, certain features of their hardware, etc.).
It actually surprises me that Linux advocates are often so narrow-minded that they cant even conceive of the possibility that some people (most home users, in fact) have very different priorities when it comes to choosing an OS. Its like the driver of a lorry deriding all those `stupid car drivers because their cars cant each carry a years supply of tomatoes.
Windows Update, of course, automatically distributes patches, so it isn't really necessary for Windows Update users to know about security holes, but theres no harm in knowing either. On the other hand, these things are generally found/patched before anyone with criminal intentions finds them. There was a similar remote exploit in every UNIX version of Netscape for literally years (all versions prior to 4.76 were vulnerable), and I haven't heard of one case of it being exploited before it was found and fixed. Of course, now that its known, anyone running pre-4.76 versions of Netscape is asking to be attacked, since, as far as I know, the only fix is to upgrade (i.e. no patch for old versions).
Xenix was abandoned because UNIX was unpopular with PC users. It was too big and slow for early PCs (UNIX was still awfully buggy in those days too), and implementing memory protection required additional memory-management hardware before the 386 (well, the 286 had a brain-damaged implementation, but Intel made clear it would be replaced in the 386).
Although Xenix was announced in 1980, it wasn't actually released until 1983 (the original IBM PC was simply incapable of running UNIX, and it took a while for machines with the necessary add-on hardware to become available). Then, after introducing the AT (which was still incapable of running a modern OS), IBM insisted on staying with the 286 for far too long, until Compaq's success forced its hand. This also held back the PC hardware architecture somewhat.
UNIX has never been nearly as popular as DOS/Win3/Win9x. Neither has WinNT/2000, although it's already surpassed UNIX in that respect. Apart from the historical reasons mentioned above, it's harder for hardware and software developers to do super-neato tricks when the OS won't let them directly access the hardware. NT and UNIX force developers to play nicely, and not all of them want to (and most home users rate features/performance of their hardware/software above stability anyway). DirectX and similar technologies were designed to solve these problems by providing efficient interfaces to certain features of the hardware, without compromising the system.
Incidentally, bash has nothing to do with UNIX. It's an enhanced clone of the old UNIX shell (the Bourne shell), but there's nothing stopping NT users running the current UNIX shell (the Korn Shell), which is freely available from AT&T (part of U/WIN, which runs on top of Win32), or even bash (via CygWin, which also runs on Win32). Since both are available in source form, they can presumably be built for Interix too, but that costs extra (though it's a native subsystem, not an emulation layer running on Win32). The problem is >90% of computer users don't want bash.
Apart from Icelandic, I cant think of any, and Iceland is very small. I think the original poster was perhaps slightly confused about the importance of Iceland.;-)
Oh rubbish. Microsoft and Intel have always been wary of each other, but continue to cooperate both with each other, and with competitors of the other. The Windows logo, for instance, did wonders to help AMD break into the x86 market, and Intels funding of Linux development is well known, as is its past UNIX work.
Believe it or not, most companies expect their business parters to also do business with their competitors. Only idiots (e.g. Linux zealots) see everything in black and white.
Do you really think so? I dont see Microsoft threatening or filing suit against all the freeware with bitmaps, icons, etc. copied straight from MS software (unlike, say, Apple, with its threats re Aqua).
I suppose its possible, but Id never let Linus Torvalds anywhere near the Windows code to begin with; not because I think hes a poor programmer, but because hes too heavily involved in developing software contaminated with the GNU General Public Virus to be of any use in commercial software work.
Why do you think this is odd? Transmeta have nothing to do with Linux. I remember the CEO visiting Microsoft back around the time the first chips were released, and giving a presentation to some of us (Im an MS developer), then chatting about Transmetas technology, product plans, etc. Transmeta recognise the PR value of having a famous guy like Linus Torvalds on their payroll, but they also know that how well their CPUs run Windows is what matters.
Since Slashdotters seem to have a problem with understanding the word `normal, Ill rephrase it as `the overwhelming majority of people. The majority of people do not program in C++, C#, VB or any other language. The majority of people, that is, do not program. If you talk to a secretary filing papers, and you suggest she create a folder called `contracts for some contracts, and one called `Contracts for others, shell think youre very strange. Such behaviour would be confusing not only to her, but to her boss, as well as any future secretary who might replace her. Similarly, imagine posting a letter to someone, only to have it returned because a letter in the address used the wrong case. Imagine if your bank refused to accept a cheque given to you because the writer used the wrong case when writing your name. Surely you recognise such things would not be acceptable.
This is a very basic UI-design concept, and the reason the Apple Lisa, Apple Macintosh, VMS, OS/2, Windows and others use case-preserving filenames is because its what people who havent been exposed to UNIX and its offshoots (such as C/C++/Java/C#) generally expect. Its a matter of custom, and the standard custom is that case is typically ignored.
The same thing applies to programming languages. Since C is case-sensitive, and has become a major programming language, many programmers (including me) prefer case-sensitive labels in our programs. However, the most popular programming language (Visual Basic) uses case-insensitive labels, as do many other languages with non-UNIX roots. When forced to deal with something in VB, I find the case-insensitivity extrememly annoying, which is similar to the reaction of most non-UNIX users to case-sensitive filenames.
I dont know why you think I said its a `big win to support variable semantics regarding case. It is quite useful for compatibility, and has nothing to do with user preference (users cannot change the behaviour). For instance, the UNIX subsystems on NT (both POSIX and Interix, the latter having been certified by TOG) provide UNIX semantics on NTFS (case-sensitive), while the Win32 subsystem provides Windows (and Macintosh, OS/2, VMS, etc.) semantics on NTFS (case-preserving). This has nothing to do with mistakes, it has to do with the behaviour expected by the overwhelming majority of users, who consider `readme and `README to be logically identical. They do not consider mistakes, such as and `readem instead of `readme to be logically identical, so youve only knocked down your own strawman in suggesting this.
Its curious that you consider it so odd that a system should be case-preserving, when its clearly the norm among non-UNIX systems. As noted above, this applies to the Apple Macintosh, IBM OS/2 and Compaq VMS, among others. Its also quite silly to be so fanatical about your operating system that youre willing to ignore obvious defects. I use UNIX and NT, and like both of them, but I recognise that both have shortcomings. Most of these are the result of legacy issues, but not all. Some of these seem perfectly reasonable to me (e.g. case sensitivity doesnt bother me in the slightest), but cause endless aggravation to others.
If you want to argue about where the case-matching should be implemented, in my view, the important level is the API level. Whether the underlying implementation is handled as part of the filesystem, within a user-mode library or in kernel-mode above the filesystem is up to the implementors. What matters at the user level is that the APIs provide what users expect, so theres consistency across applications.
Judging by the responses to my fairly simple and obvious (to normal people) comment, this topic is a sacred cow with fanatics, so Ill probably leave it at this.
You seem to think that only NT is case-preserving. In fact, most systems, including Mac OS, OS/2, VMS and Windows, use case-preserving filesystems. UNIX is the odd one, not NT.
This is why UTF-8 exists and is used on most unices, and why I consider it a better solution than plain (untrasformed) UNICODE (or UCS-16, if you like).
No it isnt. UTF-8 exists because legacy issues on old, ASCII-based systems like UNIX make the use of 2-byte UNICODE characters impracticable. Moreover, even though multi-byte schemes like UTF-8 reduce the ASCII character set to 8 bits per character, non-ASCII characters are turned into variable-length byte sequences which can be up to three bytes in length (so you dont necessarily end up with smaller strings in the non-English case, and can quite easily end up with larger ones, particularly in non-Western languages).
Apart from the length, dealing with multi-byte encoded strings is a massive pain because any given character can be between 1 and 3 bytes in size (so you cant just walk a string of unsigned bytes or shorts, as you can with ASCII and UNICODE). You need code to look at a byte, determine whether its a self-contained character or a multi-byte one, if the latter, determine how many extra bytes there are, read them in, unencode them, etc. Mind you, its better than nothing at all when youre stuck with ASCII legacy, but its no fun to use, and certainly isnt more efficient than using normal, 2-byte UNICODE (in fact its far less so).
True enough that the military are different, so if they dont trust non-German companies, its proper to avoid non-German products (not just software) in sensitive areas. However, I was responding to your comment about only fools trusting separate binary and source packages. If that was meant in a military context, I simply misunderstood.
As for me, I trust them all: my bank, my hardware suppliers, my software suppliers, etc. In fact, since Ive nothing to hide, I dont actually care if these security agencies are monitoring my every move. Theyd have to be awfully bored to be doing that, though.
Most American pension funds own shares in Microsoft, and its basic economics that there are knock-on effects. For instance, Microsoft builds buildings, buys computers, pays employees (who spend money on all manner of things), etc.
Modern economies are very interconnected, both interally and with each other. When, for instance, a Microsoft shareholder or employee buys a BMW, money flows back to Germany. You could even carry this further to an investor in the firm that builds Microsoft buildings, etc.
Bell was born in Scotland (Edinburgh), though I dont know what his citizenship status was when he developed his telephone. He may have been Canadian or American at that point.
The Zuse devices were not actually full computers, because they lacked conditional branching, so a program had only one path of execution. The link, http://www.inf.fu-berlin.de/~widiger/ICHC/papers/a nnals/node21.html, discusses early computing devices, including the first computer, the M-Mark 1, developed at the Universtity of Manchester, England. It explains why earlier machines were not universal computers.
An additional reason (according to some) why the M-Mark I qualifies as a computer, while earlier machines do not, is that it stored program code in memory. The Zuse machines, in contrast, used memory to store data, but read instructions directly from punched tape (so there was no `software, as we know it).
You may be right; the article is not very clear. I had read it as referring to two separate things: the software requirement by the German Army, and the teleconferencing concerns of the German Foreign Office. However, I can easily read it as saying the software requirement applies to both, though the teleconferencing issue clearly refers to the Foreign Office alone.
It is a number so high that it is incomprehensible, and yet may not actually represent infinity. Is this an unusual phrase? It looks quite common to me (according to Google).
Since there is a non-null intersection of the aims of NSA and Theo, and Theo does do things, there is a good argument that Theo is an agent (albeit indirect) of the NSA.
Ill put it more directly: can you prove that Theo is not a covert employee of the NSA, who is secretly putting obscure and hidden backdoors into OpenBSD (it is possible with innocent-looking source code)?
No, I cannot prove that elves haven't put backdoors into Windows.
So why the comment about Microsofts denial `proving nothing, if youre unable to prove that elves arent putting backdoors into Windows and OpenBSD? A denial can never prove anything, it can only offer an alternative explanation of something.
I'm sure that a large number of people at Microsoft have viewed some of the code, but how many have examined all the code, specifically looking for backdoors? If someone found a backdoor, whoud (s)he post the code to/.?
I have no idea what would happen if someone found a backdoor in the Windows source code, but my suspicion is that it would be reported to the owner of the code, then removed.
Taking down stale news is strange?
Actually, it is. With the cheapness of storage, most companies leave press releases and documentation up indefinitely. The Microsoft denial remains up, even though it is nearly as old as the initial Cryptonym claim. Also, this claim made Cryptonym briefly famous, and, if it were true, is something the company should be proud to have found. If I had found something as significant as an NSA backdoor in one of the most popular software products in the world, Id proudly display this fact.
If all the major news services were to run news stories (plural) about OpenBSD having a backdoor, then not only would the OpenBSD developers not deny it, but they would also deny it. ("If FALSE then TRUE" and "IF FALSE then FALSE" are both true;)
Im sorry, I dont understand. You previously said that `Microsoft would be more likely to deny the existence [of a backdoor] if there was a backdoor, yet now you say OpenBSD would also deny the existence of a backdoor in similar circumstances (your sentence is actually self-contradicting, but I think this is what you meant), namely if a baseless allegation of a backdoor were levelled at OpenBSD. By your earlier logic, this action would suggest a greater likelihood of a backdoor in OpenBSD.
I think a denial would be issued regardless of the guilt or innocence of Microsoft (or OpenBSD). The denial does not `prove anything, but the explanation provided in the Microsoft denial is far more reasonable than the original claim by Cryptonym. Moreover, the Microsoft denial/explanation remains available, while the Cryptonym one has been taken down. Surely this says something.
Lastly, to not not do X is to do X; the second `not cancels the first. If you dont believe me, ask your C compiler what !!1 is.
That is certainly reasonable, but there is no evidence of any such backdoor (the site which originally made the NSA claim has been taken down), so the whole thing is really rather ridiculous. Its no more credible than banning Intel CPUs because they may contain secret NSA instructions.
About the spying in general, it seems the Stasi had bugged all of Chancellor Kohls offices/telephones anyway, so I dont think the lazy attitude was limited to American spying. Actually, if a link in the chain is suspected of being weak, its all the more important to keep an eye on it.
Dear me, surely you know the difference between a process crashing and an operating system crashing?
Apart from the fact that its a process, not the operating system, that crashed, the result of an integer divide-by-zero is undefined, and good operating systems dont let processes continue to run after theyve generated, and not handled, fatal exceptions. I know, for instance, that FreeBSD does the same thing (core dump), and Im fairly sure Solaris does too. Systems that dont core dump can get stuck in unintentional infinite loops, which is worse, or start corrupting data (much much worse).
If your code is likely to divide by zero, and you know what you want to do in that case, you simply handle the exception. If youre not expecting it to divide by zero, and it does, it usually means a bug, so dumping core is the right thing to do (so you can either find and fix the bug, or add an exception handler).
Yes, yes, yes, but this article is discussing an apparent decision by the Bundeswehr (German Army) to ban non-German software. There is no suggestion of a general ban throughout the government.
Your standard is silly. There is an almost infinite number of things which do not exist, and for which there is insufficient proof to convince some people of this. Can you prove Theo DeRaadt is not actually an NSA agent? Can you prove that elves dont exist?
The Windows code is viewed by very large numbers of people, none of whom have backed up this claim of a `possible backdoor, which was based entirely on the name `NSAKEY (which would be an incredibly stupid name for the NSA to actually use). The most revealing thing is that the the source of this rumour has himself taken down the page discussing it. This would be a very unusual thing for the `Chief Scientist of a security firm to do about an announcement that made him famous, unless he had a very good reason. The Microsoft explanation is far more credible than the original (now deleted) announcement, so I think its obvious which is nearer to the truth.
In response to your last comment, if all the major news services were to run news stories about OpenBSD having a backdoor, do you think the OpenBSD developers would not deny it?
Either way is okay with me, but I find most people think its strange that you can have two files with the same (logical) name. Mac OS is another example of a case-preserving, but case-insensitive, OS (although, unlike NT, there is no choice in the matter), so it isnt only Windows.
Incidentally, the CompletionChar registry value is actually documented in the Microsoft KB, so its just inconvenient in comparison to most UNIX shells (but then that can be said of most aspects of CMD).
Windows 2000 handles all the non-ASCII characters I use (every day) quite well: proper sorting, command-line completion, etc. NT is all UNICODE internally, you see, including NTFS, and case mapping is a standard feature of UNICODE, so the problems faced by UNIX are not there (and this is another reason non-English support is so much better in NT, so people in small countries arent forced to use American English).
On the other hand, a UNICODE character is twice as large as an 8-bit one, so it does increase the code size somewhat. Nevertheless, it is a much better approach, and UNIX certainly would have used it if UNICODE had existed in 1969. To design an operating system post-UNICODE and use 8-bit characters would be very poor design, but UNIX has a good excuse (and most UNIXes have had UNICODE support added in some form anyway).
Yes, and they realise, for instance that in English (and most European languages), a pen is still a pen, even if its at the start of a sentence, and is therefore spelt `Pen. Apparently the designers of UNIX failed to consider that capitalisation does not generally change the meaning of a word in English (the exception being proper nouns), hence their poor design choice in this case (though in most others they chose well).
Slashdot Mirror
It actually surprises me that Linux advocates are often so narrow-minded that they cant even conceive of the possibility that some people (most home users, in fact) have very different priorities when it comes to choosing an OS. Its like the driver of a lorry deriding all those `stupid car drivers because their cars cant each carry a years supply of tomatoes.
Windows Update, of course, automatically distributes patches, so it isn't really necessary for Windows Update users to know about security holes, but theres no harm in knowing either. On the other hand, these things are generally found/patched before anyone with criminal intentions finds them. There was a similar remote exploit in every UNIX version of Netscape for literally years (all versions prior to 4.76 were vulnerable), and I haven't heard of one case of it being exploited before it was found and fixed. Of course, now that its known, anyone running pre-4.76 versions of Netscape is asking to be attacked, since, as far as I know, the only fix is to upgrade (i.e. no patch for old versions).
Although Xenix was announced in 1980, it wasn't actually released until 1983 (the original IBM PC was simply incapable of running UNIX, and it took a while for machines with the necessary add-on hardware to become available). Then, after introducing the AT (which was still incapable of running a modern OS), IBM insisted on staying with the 286 for far too long, until Compaq's success forced its hand. This also held back the PC hardware architecture somewhat.
UNIX has never been nearly as popular as DOS/Win3/Win9x. Neither has WinNT/2000, although it's already surpassed UNIX in that respect. Apart from the historical reasons mentioned above, it's harder for hardware and software developers to do super-neato tricks when the OS won't let them directly access the hardware. NT and UNIX force developers to play nicely, and not all of them want to (and most home users rate features/performance of their hardware/software above stability anyway). DirectX and similar technologies were designed to solve these problems by providing efficient interfaces to certain features of the hardware, without compromising the system.
Incidentally, bash has nothing to do with UNIX. It's an enhanced clone of the old UNIX shell (the Bourne shell), but there's nothing stopping NT users running the current UNIX shell (the Korn Shell), which is freely available from AT&T (part of U/WIN, which runs on top of Win32), or even bash (via CygWin, which also runs on Win32). Since both are available in source form, they can presumably be built for Interix too, but that costs extra (though it's a native subsystem, not an emulation layer running on Win32). The problem is >90% of computer users don't want bash.
Apart from Icelandic, I cant think of any, and Iceland is very small. I think the original poster was perhaps slightly confused about the importance of Iceland. ;-)
Believe it or not, most companies expect their business parters to also do business with their competitors. Only idiots (e.g. Linux zealots) see everything in black and white.
I suppose its possible, but Id never let Linus Torvalds anywhere near the Windows code to begin with; not because I think hes a poor programmer, but because hes too heavily involved in developing software contaminated with the GNU General Public Virus to be of any use in commercial software work.
Why do you think this is odd? Transmeta have nothing to do with Linux. I remember the CEO visiting Microsoft back around the time the first chips were released, and giving a presentation to some of us (Im an MS developer), then chatting about Transmetas technology, product plans, etc. Transmeta recognise the PR value of having a famous guy like Linus Torvalds on their payroll, but they also know that how well their CPUs run Windows is what matters.
This is a very basic UI-design concept, and the reason the Apple Lisa, Apple Macintosh, VMS, OS/2, Windows and others use case-preserving filenames is because its what people who havent been exposed to UNIX and its offshoots (such as C/C++/Java/C#) generally expect. Its a matter of custom, and the standard custom is that case is typically ignored.
The same thing applies to programming languages. Since C is case-sensitive, and has become a major programming language, many programmers (including me) prefer case-sensitive labels in our programs. However, the most popular programming language (Visual Basic) uses case-insensitive labels, as do many other languages with non-UNIX roots. When forced to deal with something in VB, I find the case-insensitivity extrememly annoying, which is similar to the reaction of most non-UNIX users to case-sensitive filenames.
I dont know why you think I said its a `big win to support variable semantics regarding case. It is quite useful for compatibility, and has nothing to do with user preference (users cannot change the behaviour). For instance, the UNIX subsystems on NT (both POSIX and Interix, the latter having been certified by TOG) provide UNIX semantics on NTFS (case-sensitive), while the Win32 subsystem provides Windows (and Macintosh, OS/2, VMS, etc.) semantics on NTFS (case-preserving). This has nothing to do with mistakes, it has to do with the behaviour expected by the overwhelming majority of users, who consider `readme and `README to be logically identical. They do not consider mistakes, such as and `readem instead of `readme to be logically identical, so youve only knocked down your own strawman in suggesting this.
Its curious that you consider it so odd that a system should be case-preserving, when its clearly the norm among non-UNIX systems. As noted above, this applies to the Apple Macintosh, IBM OS/2 and Compaq VMS, among others. Its also quite silly to be so fanatical about your operating system that youre willing to ignore obvious defects. I use UNIX and NT, and like both of them, but I recognise that both have shortcomings. Most of these are the result of legacy issues, but not all. Some of these seem perfectly reasonable to me (e.g. case sensitivity doesnt bother me in the slightest), but cause endless aggravation to others.
If you want to argue about where the case-matching should be implemented, in my view, the important level is the API level. Whether the underlying implementation is handled as part of the filesystem, within a user-mode library or in kernel-mode above the filesystem is up to the implementors. What matters at the user level is that the APIs provide what users expect, so theres consistency across applications.
Judging by the responses to my fairly simple and obvious (to normal people) comment, this topic is a sacred cow with fanatics, so Ill probably leave it at this.
Oh rubbish. NT handles case perfectly well in non-Latin scripts such Greek, Cyrillic, etc.
You seem to think that only NT is case-preserving. In fact, most systems, including Mac OS, OS/2, VMS and Windows, use case-preserving filesystems. UNIX is the odd one, not NT.
No it isnt. UTF-8 exists because legacy issues on old, ASCII-based systems like UNIX make the use of 2-byte UNICODE characters impracticable. Moreover, even though multi-byte schemes like UTF-8 reduce the ASCII character set to 8 bits per character, non-ASCII characters are turned into variable-length byte sequences which can be up to three bytes in length (so you dont necessarily end up with smaller strings in the non-English case, and can quite easily end up with larger ones, particularly in non-Western languages).
Apart from the length, dealing with multi-byte encoded strings is a massive pain because any given character can be between 1 and 3 bytes in size (so you cant just walk a string of unsigned bytes or shorts, as you can with ASCII and UNICODE). You need code to look at a byte, determine whether its a self-contained character or a multi-byte one, if the latter, determine how many extra bytes there are, read them in, unencode them, etc. Mind you, its better than nothing at all when youre stuck with ASCII legacy, but its no fun to use, and certainly isnt more efficient than using normal, 2-byte UNICODE (in fact its far less so).
As for me, I trust them all: my bank, my hardware suppliers, my software suppliers, etc. In fact, since Ive nothing to hide, I dont actually care if these security agencies are monitoring my every move. Theyd have to be awfully bored to be doing that, though.
Modern economies are very interconnected, both interally and with each other. When, for instance, a Microsoft shareholder or employee buys a BMW, money flows back to Germany. You could even carry this further to an investor in the firm that builds Microsoft buildings, etc.
Bell was born in Scotland (Edinburgh), though I dont know what his citizenship status was when he developed his telephone. He may have been Canadian or American at that point.
An additional reason (according to some) why the M-Mark I qualifies as a computer, while earlier machines do not, is that it stored program code in memory. The Zuse machines, in contrast, used memory to store data, but read instructions directly from punched tape (so there was no `software, as we know it).
You may be right; the article is not very clear. I had read it as referring to two separate things: the software requirement by the German Army, and the teleconferencing concerns of the German Foreign Office. However, I can easily read it as saying the software requirement applies to both, though the teleconferencing issue clearly refers to the Foreign Office alone.
It is a number so high that it is incomprehensible, and yet may not actually represent infinity. Is this an unusual phrase? It looks quite common to me (according to Google).
Since there is a non-null intersection of the aims of NSA and Theo, and Theo does do things, there is a good argument that Theo is an agent (albeit indirect) of the NSA.
Ill put it more directly: can you prove that Theo is not a covert employee of the NSA, who is secretly putting obscure and hidden backdoors into OpenBSD (it is possible with innocent-looking source code)?
No, I cannot prove that elves haven't put backdoors into Windows.
So why the comment about Microsofts denial `proving nothing, if youre unable to prove that elves arent putting backdoors into Windows and OpenBSD? A denial can never prove anything, it can only offer an alternative explanation of something.
I'm sure that a large number of people at Microsoft have viewed some of the code, but how many have examined all the code, specifically looking for backdoors? If someone found a backdoor, whoud (s)he post the code to /.?
I have no idea what would happen if someone found a backdoor in the Windows source code, but my suspicion is that it would be reported to the owner of the code, then removed.
Taking down stale news is strange?
Actually, it is. With the cheapness of storage, most companies leave press releases and documentation up indefinitely. The Microsoft denial remains up, even though it is nearly as old as the initial Cryptonym claim. Also, this claim made Cryptonym briefly famous, and, if it were true, is something the company should be proud to have found. If I had found something as significant as an NSA backdoor in one of the most popular software products in the world, Id proudly display this fact.
If all the major news services were to run news stories (plural) about OpenBSD having a backdoor, then not only would the OpenBSD developers not deny it, but they would also deny it. ("If FALSE then TRUE" and "IF FALSE then FALSE" are both true ;)
Im sorry, I dont understand. You previously said that `Microsoft would be more likely to deny the existence [of a backdoor] if there was a backdoor, yet now you say OpenBSD would also deny the existence of a backdoor in similar circumstances (your sentence is actually self-contradicting, but I think this is what you meant), namely if a baseless allegation of a backdoor were levelled at OpenBSD. By your earlier logic, this action would suggest a greater likelihood of a backdoor in OpenBSD.
I think a denial would be issued regardless of the guilt or innocence of Microsoft (or OpenBSD). The denial does not `prove anything, but the explanation provided in the Microsoft denial is far more reasonable than the original claim by Cryptonym. Moreover, the Microsoft denial/explanation remains available, while the Cryptonym one has been taken down. Surely this says something.
Lastly, to not not do X is to do X; the second `not cancels the first. If you dont believe me, ask your C compiler what !!1 is.
About the spying in general, it seems the Stasi had bugged all of Chancellor Kohls offices/telephones anyway, so I dont think the lazy attitude was limited to American spying. Actually, if a link in the chain is suspected of being weak, its all the more important to keep an eye on it.
Apart from the fact that its a process, not the operating system, that crashed, the result of an integer divide-by-zero is undefined, and good operating systems dont let processes continue to run after theyve generated, and not handled, fatal exceptions. I know, for instance, that FreeBSD does the same thing (core dump), and Im fairly sure Solaris does too. Systems that dont core dump can get stuck in unintentional infinite loops, which is worse, or start corrupting data (much much worse).
If your code is likely to divide by zero, and you know what you want to do in that case, you simply handle the exception. If youre not expecting it to divide by zero, and it does, it usually means a bug, so dumping core is the right thing to do (so you can either find and fix the bug, or add an exception handler).
Yes, yes, yes, but this article is discussing an apparent decision by the Bundeswehr (German Army) to ban non-German software. There is no suggestion of a general ban throughout the government.
The Windows code is viewed by very large numbers of people, none of whom have backed up this claim of a `possible backdoor, which was based entirely on the name `NSAKEY (which would be an incredibly stupid name for the NSA to actually use). The most revealing thing is that the the source of this rumour has himself taken down the page discussing it. This would be a very unusual thing for the `Chief Scientist of a security firm to do about an announcement that made him famous, unless he had a very good reason. The Microsoft explanation is far more credible than the original (now deleted) announcement, so I think its obvious which is nearer to the truth.
In response to your last comment, if all the major news services were to run news stories about OpenBSD having a backdoor, do you think the OpenBSD developers would not deny it?
Incidentally, the CompletionChar registry value is actually documented in the Microsoft KB, so its just inconvenient in comparison to most UNIX shells (but then that can be said of most aspects of CMD).
On the other hand, a UNICODE character is twice as large as an 8-bit one, so it does increase the code size somewhat. Nevertheless, it is a much better approach, and UNIX certainly would have used it if UNICODE had existed in 1969. To design an operating system post-UNICODE and use 8-bit characters would be very poor design, but UNIX has a good excuse (and most UNIXes have had UNICODE support added in some form anyway).
Did you know the Windows 2000 DDK uses command-line build tools? It's true.
Yes, and they realise, for instance that in English (and most European languages), a pen is still a pen, even if its at the start of a sentence, and is therefore spelt `Pen. Apparently the designers of UNIX failed to consider that capitalisation does not generally change the meaning of a word in English (the exception being proper nouns), hence their poor design choice in this case (though in most others they chose well).