The whole thing stems from one fundamental confusion: failing to distinguish between _viewing_ a file, and _executing_ the instructions in that file. If filetypes like VBS macros had two separate commands for these, with the default being 'view', then worms like this could not spread.
I always thought it was really stupid how the menu in Program Manager said 'Open' instead of 'Run'. Now Microsoft's decision to blur the lines between the two is coming back to haunt them.
IMHO posters to Slashdot should be able to create their own impromptu polls. So that in the post above, you'd actually be able to vote for the options and see the results, then post more comments on them.
It could get kinda stupid, but moderation will hopefully make sure that interesting or funny polls get moderated up, and stupid ones get moderated down.
What I'm thinking of is a new posting option, 'Minipoll', where you enter some XML into the comment submission box which Slashdot process and puts as a set of radio boxes inlined with your comment.
I read that the release of Netscape 6.0 will be so much like IE5 that some proprietary tags made specifically for Netscape 4 will NOT work.
My question, then, is how in the world are we supposed to maintain multi-browser compatibility?
Easy - write to the standards and only to the standards. Of course, as you point out, it's not always that simple, with people clamouring for overladen JavaScript, Flash, DHTML and the latest stupid gimmick. Just Say No:-)
There's also a pretty decent Archimedes emulator called Arcem by David Alan Gilbert, who coincidentally used to work on Amulet a few years ago. Unfortunately his site is down ATM. It's based on the GPL'ed Armulator code released by ARM Ltd. (Why is it ARM Ltd when they're a publicly traded company? )
I imagine that following the post by FascDot Killed My Pr, a lot of the posts will have a similar structure. So I'm defining a macro to avoid repetition:
#define WHY If you're such a hotshot web designer, why
Using this definition, here is my question: WHY does www.webstandards.org open links in a new browser window, when this behaviour is inconsistent with the rest of the Web, annoying, and strongly discouraged by the W3C?
Linux does not contain any Minix code, and it never did. Linus happened to be using Minix (or MINIX as it's more properly known) to start developing his Linux kernel. For convenience, he wrote code to access the Minux filesystem, to share data between the two OSes. So you could say that the Linux kernel was 'inspired by' Minix - it was kind of a reaction against it, actually - but it's not 'based on' Minix.
Maybe the answer is for a few people to read the spec, remember it, and write a new version *in their own words*. This wouldn't violate Microsoft's copyright (AFAIK; IANAL) and I don't imagine it would be possible for MS to get it banned on the grounds of being a 'trade secret'.
If somebody who has never seen Microsoft's original document nor even read the EULA they imposed on it wrote an extension to Samba (or whatever) to handle the new protocols, that would be a clean-room implementation and I imagine they'd be in the clear legally.
Emacs is the ultimate windowing system. I haven't gotten around to write a window manager in elisp yet, but I'm sure there's somebody out there that would do it.:)
Both gwm and the more recent Sawfish are extensible window managers which are implemented in an elisp-like language.
Somebody suggested, as a 'blue-sky' feature for GNOME a while ago, a kind of Emacs window manager where you have each app open in its own buffer and switch between them with C-x b, C-x o and so on. For the special case of shell buffers, we already have this. Something like a VNC-mode for XEmacs would probably allow you to implement this fairly easily.
You might be able to play Quake in wireframe or something. Or ditch the texture-mapping and just have solid coloured walls. Combine that with some simpler-looking weaponry and opponents (a big rectangular blob would be just as fearsome-looking:-) and you might get something Quake-like running at a reasonable frame rate.
It's not really a covert channel, but Mailtunnel lets you send IP packets via Internet mail messages, which is an equally twisted way of getting data from A to B.
I'm saying the different distributions can pick the best way to do feature X; they'll probably each choose something similar to what the others have, although oddballs like Red Hat and Debian might want to go their own way.
Then a consensus emerges and we end up with something fairly close to the 'best' solution. That's the way things have always worked in the past, not by having somebody decide in advance what the answer should be.
Once things have settled down, then - and only then - somebody might write a standard for it, if necessary. The reason why standards are created for network protocols is because of the need for compatibility and to avoid proprietary protocols taking over. But compatibility between different Linux distributions will mostly be enforced by the free choice of users - would you switch to the Foo distribution if it broke your software? - and there isn't a threat of proprietary, closed systems, at least not from vendors like Mandrake who keep all their code free.
This same discussion has been going on for as long as I can remember (which is not that long, but a good few years anyway).
'We need a standard to stop Linux fragmenting and becoming incompatible', etc etc. But has it happened so far? Hardly. Can you name an application which runs on one distribution but not another (for reasons other than corporate stubbornness)?
De facto standards seem to have served Linux well so far, why bother getting all formal and churning this sort of thing through endless committees?
I've seen bank tellers with such IBM keybords for desktops. Anyone know where to buy one?
pckeyboard.com bought IBM's keyboard division from Lexmark a few years ago. They still sell the IBM keyboards and a few similar models of their own. Fricking expensive, though.
A pointing device on this keyboard would be nice to have. How about some kind of 'eraserhead' pointing stick like IBM have? Or better still, you could bend and flex the keyboard to move the pointer - that would probably be great for some game, although I can't think of which game it would be.
Another cool thing would be if you could break off bits of keyboard and use them separately. For example, you could cut off the numeric keypad with a pair of scissors, or cut the keyboard in half so it's like one of those ghastly 'ergonomic' Apple things. And then if you could squish the two halves back together again...
So the kernel should display loads of messages at boot time, but if you want to have any chance of reading them, you have to run dmesg?
If so, why bother printing the messages at all? Why not just log them quietly and let the user run dmesg later?
The reason why the kernel prints messages at boot time is to help diagnose hardware problems and other serious things like failing to mount root. That's one of the reasons why booting is much more chatty than any other part of Linux - you need to see the messages immediately since you may not get the chance to run dmesg. So it's best to make sure the important messages get printed (as well as logged), and remain on screen. The less important stuff can, of course, be stored away somewhere to be inspected later.
Yeah, but it gets in the way of other more useful and informative messages.
Also, if more and more bits of kernel start displaying sponsorship messages, you'll get the problem that the messages displayed early on are scrolled off the screen within a few milliseconds. Already, even on a fairly slow machine, there are plenty of kernel messages which you never get the chance to read, because they scroll away so quickly.
It could become like MS-DOS, where every TSR or program run from autoexec.bat felt it necessary to display some sort of message. Then to make your message stand out from the rest, you had to put several blank lines above and below (Microsoft does this). Or display it in eye-catching colours, a big blue rectangle perhaps (eg Logitech's mouse driver). Finally it got to the stage where programs would clear the screen just to make sure that their useless copyright messages got noticed.
Unix was designed on the basis of not printing spurious chatty messages, so it's easy to tell when something really does require your attention. Let's keep it that way.
If failing to read a license causes me not to be bound by it, then maybe I'll just download the Linux kernel code, ignore the license, and call it public domain.
Read what the GPL says:
5. You are not required to accept this License, since you have not signed it.
You don't have to accept the GPL at all. But you will be violating copyright if you distribute Linux under any licence except the GPL.
Office2K will already save docs in a kind of bastardized HTML++ format which truly sucks because it is neither rules-following HTML nor well-formed XML
You can use HTML Tidy to correct the HTML output by Word.
While the GPL gives you certain priveledges that you wouldn't have with normal copyrighted works, it also takes some normal rights away.
ianal: It's not possible for a licence to take normal rights away (at least not without DMCA, UCITA, etc laws). The whole point of rights like fair use, parody and so on (which vary from country to country) is that they can't be taken away by the copyright holder, no matter how much crap is in the licence 'agreement'.
Now the GPL claims that a program which uses a library is a derived work of that library, but if this turns out not to be true (it's not been tested in court AFAIK) then that section of the GPL doesn't carry any weight.
In particular, it's perfectly legal for me to create a proprietary program that relies on some copyrighted shared library (.dll,.so, whatever-your-OS-uses) made by someone else.
Are you sure? Without the permission of whoever wrote the library? Have a look at the files for Microsoft Office, pick a DLL at random and distribute your own program linking with that DLL. See what happens.
But my point is that with normal copyright, I can distribute works that rely on someone else's shared library. With GPL you can't legally do this, or at least RMS hopes you can't.
IANAL, but: This issue is muddy because nobody really knows what the legal position is. There are essentially two possible cases:
Linking with a shared library does not make your program a derived work of that library. In this case, both proprietary software and GPLed software can be dynamically linked to any program you write and distribute. The GPL claims that this is not allowed, but if your program isn't a derived work of the GPLed code, then your program is not bound by the GPL's terms anyway.
Linking with a shared library does make your program a derived work. In this case, you cannot distribute a program linked with a library without permission from the library's copyright holder. Proprietary software might allow this, depending on the licence terms. Certainly something marketed as a 'library' would allow it under ceratin conditions. However, you could be on shakier ground if you wrote a program depending on 'some other company's shared library' which they hadn't given you permission to use (for example, if you wanted to use MS Word's import filters). The GPL also allows linking, treating it exactly the same as any other derived work - ie if you distribute your program, you must distribute it only under the GPL.
Slashdot Mirror
The whole thing stems from one fundamental confusion: failing to distinguish between _viewing_ a file, and _executing_ the instructions in that file. If filetypes like VBS macros had two separate commands for these, with the default being 'view', then worms like this could not spread.
I always thought it was really stupid how the menu in Program Manager said 'Open' instead of 'Run'. Now Microsoft's decision to blur the lines between the two is coming back to haunt them.
IMHO posters to Slashdot should be able to create their own impromptu polls. So that in the post above, you'd actually be able to vote for the options and see the results, then post more comments on them.
It could get kinda stupid, but moderation will hopefully make sure that interesting or funny polls get moderated up, and stupid ones get moderated down.
What I'm thinking of is a new posting option, 'Minipoll', where you enter some XML into the comment submission box which Slashdot process and puts as a set of radio boxes inlined with your comment.
I think you have a spurious comma after 'It'.
Easy - write to the standards and only to the standards. Of course, as you point out, it's not always that simple, with people clamouring for overladen JavaScript, Flash, DHTML and the latest stupid gimmick. Just Say No :-)
There's also a pretty decent Archimedes emulator called Arcem by David Alan Gilbert, who coincidentally used to work on Amulet a few years ago. Unfortunately his site is down ATM. It's based on the GPL'ed Armulator code released by ARM Ltd. (Why is it ARM Ltd when they're a publicly traded company? )
I imagine that following the post by FascDot Killed My Pr, a lot of the posts will have a similar structure. So I'm defining a macro to avoid repetition:
#define WHY If you're such a hotshot web designer, whyUsing this definition, here is my question:
WHY does www.webstandards.org open links in a new browser window, when this behaviour is inconsistent with the rest of the Web, annoying, and strongly discouraged by the W3C?
Linux does not contain any Minix code, and it never did. Linus happened to be using Minix (or MINIX as it's more properly known) to start developing his Linux kernel. For convenience, he wrote code to access the Minux filesystem, to share data between the two OSes. So you could say that the Linux kernel was 'inspired by' Minix - it was kind of a reaction against it, actually - but it's not 'based on' Minix.
Maybe the answer is for a few people to read the spec, remember it, and write a new version *in their own words*. This wouldn't violate Microsoft's copyright (AFAIK; IANAL) and I don't imagine it would be possible for MS to get it banned on the grounds of being a 'trade secret'.
If somebody who has never seen Microsoft's original document nor even read the EULA they imposed on it wrote an extension to Samba (or whatever) to handle the new protocols, that would be a clean-room implementation and I imagine they'd be in the clear legally.
Both gwm and the more recent Sawfish are extensible window managers which are implemented in an elisp-like language.
Somebody suggested, as a 'blue-sky' feature for GNOME a while ago, a kind of Emacs window manager where you have each app open in its own buffer and switch between them with C-x b, C-x o and so on. For the special case of shell buffers, we already have this. Something like a VNC-mode for XEmacs would probably allow you to implement this fairly easily.
You might be able to play Quake in wireframe or something. Or ditch the texture-mapping and just have solid coloured walls. Combine that with some simpler-looking weaponry and opponents (a big rectangular blob would be just as fearsome-looking :-) and you might get something Quake-like running at a reasonable frame rate.
It's not really a covert channel, but Mailtunnel lets you send IP packets via Internet mail messages, which is an equally twisted way of getting data from A to B.
I'm saying the different distributions can pick the best way to do feature X; they'll probably each choose something similar to what the others have, although oddballs like Red Hat and Debian might want to go their own way.
Then a consensus emerges and we end up with something fairly close to the 'best' solution. That's the way things have always worked in the past, not by having somebody decide in advance what the answer should be.
Once things have settled down, then - and only then - somebody might write a standard for it, if necessary. The reason why standards are created for network protocols is because of the need for compatibility and to avoid proprietary protocols taking over. But compatibility between different Linux distributions will mostly be enforced by the free choice of users - would you switch to the Foo distribution if it broke your software? - and there isn't a threat of proprietary, closed systems, at least not from vendors like Mandrake who keep all their code free.
This same discussion has been going on for as long as I can remember (which is not that long, but a good few years anyway).
'We need a standard to stop Linux fragmenting and becoming incompatible', etc etc. But has it happened so far? Hardly. Can you name an application which runs on one distribution but not another (for reasons other than corporate stubbornness)?
De facto standards seem to have served Linux well so far, why bother getting all formal and churning this sort of thing through endless committees?
pckeyboard.com bought IBM's keyboard division from Lexmark a few years ago. They still sell the IBM keyboards and a few similar models of their own. Fricking expensive, though.
A pointing device on this keyboard would be nice to have. How about some kind of 'eraserhead' pointing stick like IBM have? Or better still, you could bend and flex the keyboard to move the pointer - that would probably be great for some game, although I can't think of which game it would be.
Another cool thing would be if you could break off bits of keyboard and use them separately. For example, you could cut off the numeric keypad with a pair of scissors, or cut the keyboard in half so it's like one of those ghastly 'ergonomic' Apple things. And then if you could squish the two halves back together again...
I quite agree. If you get a message that says "If you get an email that says 'I love you', DON'T OPEN IT!", don't open it!
Yeah, and you wouldn't get to know what the box is doing if the information is crowded off the screen by lots of advertising.
A very sharp cookie? Sounds painful.
So the kernel should display loads of messages at boot time, but if you want to have any chance of reading them, you have to run dmesg?
If so, why bother printing the messages at all? Why not just log them quietly and let the user run dmesg later?
The reason why the kernel prints messages at boot time is to help diagnose hardware problems and other serious things like failing to mount root. That's one of the reasons why booting is much more chatty than any other part of Linux - you need to see the messages immediately since you may not get the chance to run dmesg. So it's best to make sure the important messages get printed (as well as logged), and remain on screen. The less important stuff can, of course, be stored away somewhere to be inspected later.
Truly elite programmers would write a Linus simulator, feed it parameters corresponding to Finland in 1990, and let it start coding.
Yeah, but it gets in the way of other more useful and informative messages.
Also, if more and more bits of kernel start displaying sponsorship messages, you'll get the problem that the messages displayed early on are scrolled off the screen within a few milliseconds. Already, even on a fairly slow machine, there are plenty of kernel messages which you never get the chance to read, because they scroll away so quickly.
It could become like MS-DOS, where every TSR or program run from autoexec.bat felt it necessary to display some sort of message. Then to make your message stand out from the rest, you had to put several blank lines above and below (Microsoft does this). Or display it in eye-catching colours, a big blue rectangle perhaps (eg Logitech's mouse driver). Finally it got to the stage where programs would clear the screen just to make sure that their useless copyright messages got noticed.
Unix was designed on the basis of not printing spurious chatty messages, so it's easy to tell when something really does require your attention. Let's keep it that way.
You don't have to accept the GPL at all. But you will be violating copyright if you distribute Linux under any licence except the GPL.
You can use HTML Tidy to correct the HTML output by Word.
ianal: It's not possible for a licence to take normal rights away (at least not without DMCA, UCITA, etc laws). The whole point of rights like fair use, parody and so on (which vary from country to country) is that they can't be taken away by the copyright holder, no matter how much crap is in the licence 'agreement'.
Now the GPL claims that a program which uses a library is a derived work of that library, but if this turns out not to be true (it's not been tested in court AFAIK) then that section of the GPL doesn't carry any weight.
Are you sure? Without the permission of whoever wrote the library? Have a look at the files for Microsoft Office, pick a DLL at random and distribute your own program linking with that DLL. See what happens.
IANAL, but: This issue is muddy because nobody really knows what the legal position is. There are essentially two possible cases: