I think a lot of young people who write open source are aiming for a public domain or BSD license,
Honestly, when I think of a license, it tends to be "Give me credit, money, and sexy women." And then I go "Okay, nobody's going to like that license; how about I go with the GPL, which prevents people from blatently ripping me off." I think a lot of young programmers want to use a harsh nasty license, but settle on the GPL as it's acceptable and cool in the open source world.
The point is not duplicating simple things that would save everyone time.
But you make us duplicate your simple things.
Software could have gotten a lot further a lot faster if there had been an "reference" UNIX spec
If you want to make a reference spec, form a consortium and make it. Somehow, I don't feel like making a reference spec so you can take it, improve it and sell it back to me. I want to build the best system on the market, not the least common denomenator that everyone rips off. "Cool, you built an Unix kernel. But if you want to run it on a Sun, you have to buy our Unix system, which is a rip off of yours." Not my game.
One of the inventors of Unix is wondering why the GNU (and by extension Linux) community is rebuilding something he made 30 years ago. I've been wondering the same thing myself. Aren't there any better ideas in the past 3 decades?
Thirty years ago, they didn't have vi or emacs. Thirty years ago, they didn't have WYSIWYG wordprocessors. Thirty years ago, they didn't have C++ or Perl or Python. Thirty years ago, they used ASCII, and maybe a few hacks that replaced ASCII punctuation with accented characters.
Rebuilding what he made 30 years, on a modern system, could be done by a dozen undergrads in a semester. You aren't running a system they had 30 years. You're running a massively evolved version of that system that continues to evolve.
how is Taco going to feel when he reads about a Chinook full of young Americans that got shot down from a SAM battery that could not be hacked and disabled because Achmed reads Slashdot,
Right. The Iraq government can afford to run a net connection to every SAM battery, and is stupid enough to leave it open to the Internet. Even societies that can afford to network their military, don't connect the Internet to internal systems - it's just stupid. The most secure connection is none at all, and everyone knows that.
I'm sure it's an unpopular view amongst the freedom of IP at all costs crowd that's common here, but maybe for the duration of the Iraqi conflict, we can stop posting exploit and bug notifications, at least until the US has installed a nascent capitalist, western ideologued democracy in Iraq.
I didn't realize we were planning on slaughtering all the Iraqs and colonizing the country. That, of course, is about the only way we're going to get a "western ideologued" society in Iraq. What do you think the odds are that we can even get some sort of stable democracy going - historically, "our son of a bitch" governments have been common - and it's questionable whether you can just stick a democracy in a country that has no concept of one and have it thrive.
It's about the same age as VS 6.0, then, but I don't see anyone stopping ragging on it:P
You mean imacat writing
For all this trouble, you get a compiler that isn't much improved. It still tells you to call Microsoft customer support at the first non-trivial use of C++. It still doesn't like non-inline template specialization.
talking about VC++.NET, or antity's response of
Now read that "1999" bit again. How, exactly, did "standard" C++ look back about 3.5 years ago?!
What did you expect? Hell, many modern C++ features weren't even implemented by most available compilers in 1999.
?
In any case, who cares what people are ragging on? The complaining about GCC 2.95's standard compliance is absurd, whether or not any other complaints are absurd.
But ease of use? I installed a DNS server the other night (djbdns - very cool app), but it's simple 5 step install process took me about 4 attempts and 2 hours to get working right (including configuration).
djb is an ass, who doesn't give a damn how hard it is for you to install his software. Are you going to tell me that there's no software that's hard to install on Windows? Install bind from a package and then discuss how hard it is to install software on Linux.
Re:I'm confused.
on
Baked Apple
·
· Score: 2, Insightful
And she dumped it, without explanation, at a repair shop, because fixing it would be $1,000.00 for the new screen?
They said it's going to be at least a thousand dollars for the screen alone. And after you've poured another couple thousand dollars into the thing, you've got a laptop that's been baked in an overn, and likely never works quite right. It's almost cheaper, and certainly easier to just replace the whole thing.
Even GCC falls short of 100% ANSI compliance. (Ever try to call the stream manipulator "fixed" in GCC 2.9x? It's an exercise in futility, because you simply won't find it.)
True, GCC is not 100% ISO compliant, but it's really not fair to bring up stuff about 2.9x. 2.95 was released almost five years ago, and 3.0 (its much more compliant successor) was released almost three years ago, with the even more compliant 3.1 (aka 3.2) being almost two years old.
The Linux community will be much better off when geeks realize that there is no difference between a geek and a consumer.
Bull. To use your car analogy; I, the average consumer, would never drive a mustang. Does that mean that there aren't car geeks out there who love mustangs and want the power? Of course not.
I don't use Windows, because it doesn't do what I f**** tell it to, quietly and efficently. I use Linux because I can quickly write a shell script to input my scans and format them just the way I like them, and then I can run netpbm over them to get them to just the way my ocr program likes them. Random windows don't pop up; it doesn't ask me if I'm sure every time I delete a file. I use my computer to scan books and process those scans for Project Gutenberg.
These types of geeks aren't qualified to determine what should be in an OS since they don't even know for themselves what they are using their computers for
They know what they're using thier computers for. Just because your toys come in different packages, doesn't make theirs less valid.
There's nothing that stops someone from whipping open extra CD #7 and installing the obscure browser and mp3 player they like.
If you're installing a multi-gigabyte OS, the obscure browser and mp3 player are going to be on CD #7. If you don't want to use it, then don't, but don't complain that when you do a full installation with all 12 CD's, you got just what you asked for.
If we had one standard packaging mechanism, standardized desktops, and windowing environments, Linux would be doing much better.
We did have a standardized windowing enviroment: it was called CDE, and was generally hated.
Diversity builds strength. How many advances were made only because Red Hat or Debian or Mandrake could do it, so we had to do it? Gnome and KDE have been playing off each other quite nicely, and allow the exploration of different routes. I run Linux because there is choice, and because there isn't one true way. I suspect many of my fellow Linux desktop users run Linux because they can adjust it to be exactly what they want. Throwing out your current market to get another market is almost always a bad idea.
Linux will be much weaker than [...] BEOS in desktop strength.
Weaker than BeOS? In what sense? BeOS is dead, because it couldn't get a market. Linux works for more people as a desktop then BeOS ever did.
I dont see these layers standardizing anytime soon, and they will divide the precious developer pool.
Divide the precious developer pool? There are some 400 developers in Debian, and we can't use more, even if Red Hat and Mandrake all disbanded and made Debian the one true distribution. Konqueror and Mozilla build each other up; as each one adds an great innovation, it gets processed, the problems fixed, and added to the other. That wouldn't happen if all the people were working on one browser.
In any case, Microsoft and MacOS ignore the developer pool; if they aren't paying you, you aren't touching anything of any importance. There's always a place for a new developer in Linux, and if it's reinventing the wheel, well, I doubt you still drive on crude stone wheels.
Re:Honest comparison between Gnome and KDE?
on
Gnome 2.2 Released
·
· Score: 1
The main difference is that KDE's file manager is average, and gnome's is below average. (so people seem to agree)
I've used both - more of Nautilus - and I don't agree at all. Konqueror is strictly average; being a console jockey, I didn't find much use for it. Nautilus is much better at previews and other slow stuff; it was slow and painful, but I rearranged my graphics directory using it, because it beat trying to do so by the command line.
I was forming my own opinions instead of abrogating this responsibility to the television.
At one point in my life, I found myself crossing myself. I'm not Catholic, or even close to that. It just happened to be what some fictional characters did. The fact is, though, they weren't TV characters; they were literary characters. It would have been even easier to pick up those traits if they were real people.
Sure, if you isolate yourself from all outside influence, you won't pick up traits from that influence. You also won't pick up entertainment or information. And this goes for all forms of influence. One of the big reasons you're anti-war in Iraq is probably because your influences are anti-war. If you still watched no TV, but read the National Review and Veterans Today and went to the meetings of your local milita group, who thought Iraq would be excellent for training, you too could believe full heartedly in the war.
There has been several good movies recently (and many not-so-recently) remaking Shakespeare, who ripped off the stories of many authors preceding him. I don't remember any movies remaking works of Pound.
Furthermore, what I remember of Pound's works were for their unusual form and our tortured analysis of them. Of Shakespeare's works, I remember dramatic characters and events, working out timeless events of love and revenge.
There are two rules that Pound forgot - "There's nothing new under the sun", and "The audience doesn't want something unique - they want something they can connect to."
But not necessarily highest. The Russian alphabet scores higher - most case pairs in Russian look the same, unlike English, and has a few more characters, but not too many.
IP addresses change.
So do domain names; many domain names are linked to IP addresses. (I was once x8b4e53cd.dhcp.okstate.edu.)
my lawyers won't be able to tell one Asian warez site from another!
If you're sueing Asian warez sites, you better get an Asian lawyer, or he's going to have a hard time in court!
I still think there is an awful lot to be said for getting on with the little irks in life as opposed to insisting they be stamped out.
And an awful lot to be said for the fact that life has a huge number of little irks, so we may as well smooth out those we can.
If we're only ever provided with what we already know we like, how can we ever change?
A lot of people have already seem the movie and just want to show it to their children. Honestly, most of us have heard the cuss words and seen the blood and nudity; I doubt a movie is going to convince us to like them any more. In one case mentioned in this discussion, the person has heard way too many gun shots and knows what that the sounds of war will do to him. What advantage does he gain by listening to them for the a millionth time?
However, it is a terrible sign of the instant-gratification-guaranteed culture that we should want to watch them in any way other than that which was intended.
Um, editing works of art to fit what you want has been around for a long time. As the link points out, Bowdler's Family Shakespeare is classic, almost exactly the same as what's being done here, and happened in 1818, almost 200 years before our current "instant-gratification-guaranteed culture".
Actually, the best thing to do would be to create a new unification system for alphabetic scripts, matching the Han unification for Chinese-based scripts, just for use with URLs.
Actually, they do use the same unification system for alphabet scripts as they do for the Chinese script.
That way Cyrillic o, omicron, and Latin o would be the same DNS codepoint
What about a and , or and Y? And what about the Armenian ?
If we don't permit case-sensitive & don't allow wacky ASCII characters, why should we allow all sixty or so variations on the character 'a' that are recognised by Unicode?
So that people can use names that are meaningful to them, instead of having to mangle the name of their organization into a very limited foreign character set? 0/O, and l/I/1 are allowed in ASCII (both paypal and paypai), so it's not like confusion didn't already potentially exist.
Not only is it the shortest, simplest character set the world has to offer,
Not. 0 and 1 is shorter and simpler, as is the Latin alphabet.
But if DNS hostnames start to come in Kanji or Hangul, it will be inestimably worse.
In this day and age, most of us have the ability to cut and paste. Using a kanji or a hangul domain name is a sign your targeted audience has no problem with kanji or hangul.
It's trivial to print the whole English alphabet on a single page, and with a rudimentary pronounciation-guide too.
Not really a useful one, considering the odd digraphs and the fact that DNS names aren't necessarily English.
How'd you like that as the hostname that's been DNSing you? Try reading it over the phone to the upstream sysadmin, maybe?
How do you pronounce ztz01588a.xxqcji.org? That's a valid domain name. Try reading the IP address.
every literate human can reasonably read, input, and maybe even pronounce.
Knowledge of the English alphabet is not required for literacy.
To handle single-byte international character sets wouldn't have been that difficult
Except for the fact that there's no such thing; a single-byte character set can cover common usage for western Europe, or Greece, but not both, and there's no single-byte character set that can cover Japanese or Chinese.
But Unicode's two-byte characters make this fail badly - if the bytes happen to be aa, changing them to AA gets you an entirely unrelated character, and vice versa
They aren't just sticking in UTF-16 and hoping it works. aa is aa, not a double byte character.
But on the other hand, a big point of the internet is that it's supposed to be international, how are for instance americans supposed to type unique swedish characters to find the web site?
If your audience is international, then you chose a domain name that they all can type. If your webpage is only in Armenian, it doesn't really matter if your domain name is too; if someone else wants to read it, they can cut and paste the link in.
Does anyone else think the Gutenberg project is being incredibly hypocritical with their redistribution rules and stuff?
If you use our trademark, don't screw it up? I don't see why that's such a hypocritical license. Just remove the header and don't use the PG name, and you'll be all right.
I doubt they would have any legal ground to sue you.
It's a trademark license; I don't see any reason why they can't stop you from distributing stuff marked "Project Gutenberg", no matter what it is.
miscompiled code? it should cause compiler warnings/errors. otherwise it IS compiler's fault.
A compiler's not magic. It can't detect all code which has illegal aliasing. (It's equivalent to detecting whether a Turing machine will return a value.) So it will probably produce miscompiled code for some code that's illegal. Still your bug. Even ICE's on non-complaint code are legal, though not friendly.
Slashdot Mirror
I think a lot of young people who write open source are aiming for a public domain or BSD license,
Honestly, when I think of a license, it tends to be "Give me credit, money, and sexy women." And then I go "Okay, nobody's going to like that license; how about I go with the GPL, which prevents people from blatently ripping me off." I think a lot of young programmers want to use a harsh nasty license, but settle on the GPL as it's acceptable and cool in the open source world.
The point is not duplicating simple things that would save everyone time.
But you make us duplicate your simple things.
Software could have gotten a lot further a lot faster if there had been an "reference" UNIX spec
If you want to make a reference spec, form a consortium and make it. Somehow, I don't feel like making a reference spec so you can take it, improve it and sell it back to me. I want to build the best system on the market, not the least common denomenator that everyone rips off. "Cool, you built an Unix kernel. But if you want to run it on a Sun, you have to buy our Unix system, which is a rip off of yours." Not my game.
One of the inventors of Unix is wondering why the GNU (and by extension Linux) community is rebuilding something he made 30 years ago. I've been wondering the same thing myself. Aren't there any better ideas in the past 3 decades?
Thirty years ago, they didn't have vi or emacs. Thirty years ago, they didn't have WYSIWYG wordprocessors. Thirty years ago, they didn't have C++ or Perl or Python. Thirty years ago, they used ASCII, and maybe a few hacks that replaced ASCII punctuation with accented characters.
Rebuilding what he made 30 years, on a modern system, could be done by a dozen undergrads in a semester. You aren't running a system they had 30 years. You're running a massively evolved version of that system that continues to evolve.
how is Taco going to feel when he reads about a Chinook full of young Americans that got shot down from a SAM battery that could not be hacked and disabled because Achmed reads Slashdot,
Right. The Iraq government can afford to run a net connection to every SAM battery, and is stupid enough to leave it open to the Internet. Even societies that can afford to network their military, don't connect the Internet to internal systems - it's just stupid. The most secure connection is none at all, and everyone knows that.
I'm sure it's an unpopular view amongst the freedom of IP at all costs crowd that's common here, but maybe for the duration of the Iraqi conflict, we can stop posting exploit and bug notifications, at least until the US has installed a nascent capitalist, western ideologued democracy in Iraq.
I didn't realize we were planning on slaughtering all the Iraqs and colonizing the country. That, of course, is about the only way we're going to get a "western ideologued" society in Iraq. What do you think the odds are that we can even get some sort of stable democracy going - historically, "our son of a bitch" governments have been common - and it's questionable whether you can just stick a democracy in a country that has no concept of one and have it thrive.
You mean imacat writing
talking about VC++.NET, or antity's response of
?
In any case, who cares what people are ragging on? The complaining about GCC 2.95's standard compliance is absurd, whether or not any other complaints are absurd.
But ease of use? I installed a DNS server the other night (djbdns - very cool app), but it's simple 5 step install process took me about 4 attempts and 2 hours to get working right (including configuration).
djb is an ass, who doesn't give a damn how hard it is for you to install his software. Are you going to tell me that there's no software that's hard to install on Windows? Install bind from a package and then discuss how hard it is to install software on Linux.
And she dumped it, without explanation, at a repair shop, because fixing it would be $1,000.00 for the new screen?
They said it's going to be at least a thousand dollars for the screen alone. And after you've poured another couple thousand dollars into the thing, you've got a laptop that's been baked in an overn, and likely never works quite right. It's almost cheaper, and certainly easier to just replace the whole thing.
Even GCC falls short of 100% ANSI compliance. (Ever try to call the stream manipulator "fixed" in GCC 2.9x? It's an exercise in futility, because you simply won't find it.)
True, GCC is not 100% ISO compliant, but it's really not fair to bring up stuff about 2.9x. 2.95 was released almost five years ago, and 3.0 (its much more compliant successor) was released almost three years ago, with the even more compliant 3.1 (aka 3.2) being almost two years old.
The Linux community will be much better off when geeks realize that there is no difference between a geek and a consumer.
Bull. To use your car analogy; I, the average consumer, would never drive a mustang. Does that mean that there aren't car geeks out there who love mustangs and want the power? Of course not.
I don't use Windows, because it doesn't do what I f**** tell it to, quietly and efficently. I use Linux because I can quickly write a shell script to input my scans and format them just the way I like them, and then I can run netpbm over them to get them to just the way my ocr program likes them. Random windows don't pop up; it doesn't ask me if I'm sure every time I delete a file. I use my computer to scan books and process those scans for Project Gutenberg.
These types of geeks aren't qualified to determine what should be in an OS since they don't even know for themselves what they are using their computers for
They know what they're using thier computers for. Just because your toys come in different packages, doesn't make theirs less valid.
There's nothing that stops someone from whipping open extra CD #7 and installing the obscure browser and mp3 player they like.
If you're installing a multi-gigabyte OS, the obscure browser and mp3 player are going to be on CD #7. If you don't want to use it, then don't, but don't complain that when you do a full installation with all 12 CD's, you got just what you asked for.
If we had one standard packaging mechanism, standardized desktops, and windowing environments, Linux would be doing much better.
We did have a standardized windowing enviroment: it was called CDE, and was generally hated.
Diversity builds strength. How many advances were made only because Red Hat or Debian or Mandrake could do it, so we had to do it? Gnome and KDE have been playing off each other quite nicely, and allow the exploration of different routes. I run Linux because there is choice, and because there isn't one true way. I suspect many of my fellow Linux desktop users run Linux because they can adjust it to be exactly what they want. Throwing out your current market to get another market is almost always a bad idea.
Linux will be much weaker than [...] BEOS in desktop strength.
Weaker than BeOS? In what sense? BeOS is dead, because it couldn't get a market. Linux works for more people as a desktop then BeOS ever did.
I dont see these layers standardizing anytime soon, and they will divide the precious developer pool.
Divide the precious developer pool? There are some 400 developers in Debian, and we can't use more, even if Red Hat and Mandrake all disbanded and made Debian the one true distribution. Konqueror and Mozilla build each other up; as each one adds an great innovation, it gets processed, the problems fixed, and added to the other. That wouldn't happen if all the people were working on one browser.
In any case, Microsoft and MacOS ignore the developer pool; if they aren't paying you, you aren't touching anything of any importance. There's always a place for a new developer in Linux, and if it's reinventing the wheel, well, I doubt you still drive on crude stone wheels.
The main difference is that KDE's file manager is average, and gnome's is below average. (so people seem to agree)
I've used both - more of Nautilus - and I don't agree at all. Konqueror is strictly average; being a console jockey, I didn't find much use for it. Nautilus is much better at previews and other slow stuff; it was slow and painful, but I rearranged my graphics directory using it, because it beat trying to do so by the command line.
Note SERIAL NUMBERS AND CD'S ARE NOT PROOF OF PURCHASE. You MUST HAVE RECEIPTS!
Whatever happened to possession being 9/10ths of the law? I have the CD's; I have no more obligation to keep receipts then I do on the chairs I have.
I was forming my own opinions instead of abrogating this responsibility to the television.
At one point in my life, I found myself crossing myself. I'm not Catholic, or even close to that. It just happened to be what some fictional characters did. The fact is, though, they weren't TV characters; they were literary characters. It would have been even easier to pick up those traits if they were real people.
Sure, if you isolate yourself from all outside influence, you won't pick up traits from that influence. You also won't pick up entertainment or information. And this goes for all forms of influence. One of the big reasons you're anti-war in Iraq is probably because your influences are anti-war. If you still watched no TV, but read the National Review and Veterans Today and went to the meetings of your local milita group, who thought Iraq would be excellent for training, you too could believe full heartedly in the war.
We all know what Ezra Pound preached right?
Make it new.
There has been several good movies recently (and many not-so-recently) remaking Shakespeare, who ripped off the stories of many authors preceding him. I don't remember any movies remaking works of Pound.
Furthermore, what I remember of Pound's works were for their unusual form and our tortured analysis of them. Of Shakespeare's works, I remember dramatic characters and events, working out timeless events of love and revenge.
There are two rules that Pound forgot - "There's nothing new under the sun", and "The audience doesn't want something unique - they want something they can connect to."
What if you're looking for the abuse@ address for a Chinese ISP that host spammers that spam you in Chinese?
That's what cut and paste was made for.
The English alphabet scores high on all scales.
But not necessarily highest. The Russian alphabet scores higher - most case pairs in Russian look the same, unlike English, and has a few more characters, but not too many.
IP addresses change.
So do domain names; many domain names are linked to IP addresses. (I was once x8b4e53cd.dhcp.okstate.edu.)
my lawyers won't be able to tell one Asian warez site from another!
If you're sueing Asian warez sites, you better get an Asian lawyer, or he's going to have a hard time in court!
I still think there is an awful lot to be said for getting on with the little irks in life as opposed to insisting they be stamped out.
And an awful lot to be said for the fact that life has a huge number of little irks, so we may as well smooth out those we can.
If we're only ever provided with what we already know we like, how can we ever change?
A lot of people have already seem the movie and just want to show it to their children. Honestly, most of us have heard the cuss words and seen the blood and nudity; I doubt a movie is going to convince us to like them any more. In one case mentioned in this discussion, the person has heard way too many gun shots and knows what that the sounds of war will do to him. What advantage does he gain by listening to them for the a millionth time?
However, it is a terrible sign of the instant-gratification-guaranteed culture that we should want to watch them in any way other than that which was intended.
Um, editing works of art to fit what you want has been around for a long time. As the link points out, Bowdler's Family Shakespeare is classic, almost exactly the same as what's being done here, and happened in 1818, almost 200 years before our current "instant-gratification-guaranteed culture".
Actually, the best thing to do would be to create a new unification system for alphabetic scripts, matching the Han unification for Chinese-based scripts, just for use with URLs.
Actually, they do use the same unification system for alphabet scripts as they do for the Chinese script.
That way Cyrillic o, omicron, and Latin o would be the same DNS codepoint
What about a and , or and Y? And what about the Armenian ?
If we don't permit case-sensitive & don't allow wacky ASCII characters, why should we allow all sixty or so variations on the character 'a' that are recognised by Unicode?
So that people can use names that are meaningful to them, instead of having to mangle the name of their organization into a very limited foreign character set? 0/O, and l/I/1 are allowed in ASCII (both paypal and paypai), so it's not like confusion didn't already potentially exist.
Not only is it the shortest, simplest character set the world has to offer,
Not. 0 and 1 is shorter and simpler, as is the Latin alphabet.
But if DNS hostnames start to come in Kanji or Hangul, it will be inestimably worse.
In this day and age, most of us have the ability to cut and paste. Using a kanji or a hangul domain name is a sign your targeted audience has no problem with kanji or hangul.
It's trivial to print the whole English alphabet on a single page, and with a rudimentary pronounciation-guide too.
Not really a useful one, considering the odd digraphs and the fact that DNS names aren't necessarily English.
How'd you like that as the hostname that's been DNSing you? Try reading it over the phone to the upstream sysadmin, maybe?
How do you pronounce ztz01588a.xxqcji.org? That's a valid domain name. Try reading the IP address.
every literate human can reasonably read, input, and maybe even pronounce.
Knowledge of the English alphabet is not required for literacy.
Is one russian charset has been chosen over others?
Yes. An encoding of Unicode is used, because that can handle every language in the world and needs no language tagging.
To handle single-byte international character sets wouldn't have been that difficult
Except for the fact that there's no such thing; a single-byte character set can cover common usage for western Europe, or Greece, but not both, and there's no single-byte character set that can cover Japanese or Chinese.
But Unicode's two-byte characters make this fail badly - if the bytes happen to be aa, changing them to AA gets you an entirely unrelated character, and vice versa
They aren't just sticking in UTF-16 and hoping it works. aa is aa, not a double byte character.
But on the other hand, a big point of the internet is that it's supposed to be international,
how are for instance americans supposed to type unique swedish characters to find the web site?
If your audience is international, then you chose a domain name that they all can type. If your webpage is only in Armenian, it doesn't really matter if your domain name is too; if someone else wants to read it, they can cut and paste the link in.
Does anyone else think the Gutenberg project is being incredibly hypocritical with their redistribution rules and stuff?
If you use our trademark, don't screw it up? I don't see why that's such a hypocritical license. Just remove the header and don't use the PG name, and you'll be all right.
I doubt they would have any legal ground to sue you.
It's a trademark license; I don't see any reason why they can't stop you from distributing stuff marked "Project Gutenberg", no matter what it is.
miscompiled code? it should cause compiler warnings/errors. otherwise it IS compiler's fault.
A compiler's not magic. It can't detect all code which has illegal aliasing. (It's equivalent to detecting whether a Turing machine will return a value.) So it will probably produce miscompiled code for some code that's illegal. Still your bug. Even ICE's on non-complaint code are legal, though not friendly.