Err. I hate to say it - no, really, I do, but when I talk to people that don't run Linux, they've either heard of Red Hat, Mandrake, or SuSE.
Also, this may be just because I don't seek out GUI interfaces to stuff, but when I ran FreeBSD 4.5, to install software you had to go to a CLI, type in the install command... and wait for it to compile. The article "for people new to FreeBSD and Unix" on freebsd.org seems to verify this. And of course, FreeBSD uses a text-mode installer. Based just on *that*, FreeBSD wouldn't fit her criteria. (Although I did like it when I used it.)
My point is, her technique was probably closest to what someone clueless, yet still brave enough to go try it, would actually do, which is what she claimed to be doing. So, criticizing her for "let's see, not enough variety" just doesn't seem right. I don't know how many people *about* to start out with Linux would have even heard of NetBSD before.
Bottom line: never upgrade kernels immediately on ANY production machine. (Never do it unless you absolutely have to, in fact.) Case in point: 2.4.18 through 2.4.20 released with bugs in ext3, notably the journal=data bug.
Processing an icon takes another level of brain processing, another level of indirection.
You've never played DDR, or maybe even video games, have you? When you're familiar with it, processing an image is pretty much instantaneous, not that it's terribly slow in the first place for simple stuff like this.
This provides asegwayinto using CVS for controlling revision of large software projects.
... the word you're looking for up there is segué. Although that would be kind of cool... hey, I set my CVS repository up a couple weeks ago! Do I get a Segway now?;_;
I think our friend Al Viro would have something to say about the auditing level of the Linux kernel. And if we're talking about drivers/ in particular, it would probably involve the words "obfuscated", "brain dead", "steaming pile of shit", "warped beyond all belief"...:)
Linux code gets a fair amount of review. But once it's there, there really isn't any auditing at all.
Interesting choice - apparently GCC 3.3 includes a lot of work SuSE have contributed. Will this be as controversial as Redhat's compiler choice of 2.96 a while back?
Probably not. GCC 3.3 will be an actual official FSF release. By the time 12 April comes around I doubt SuSE's gcc lags behind the official FSF release much. Remember, 2.96 and 2.97 were basically continuations of the 2.9x branch, after the FSF had basically stopped working on it and started working on GCC 3.x.
Thing is, gcc 3.x broke things. Also, Red Hat had a collection of IA-64 improvements for gcc that may not have made it into mainline yet. So, they made the unofficial releases because they felt that's what served their customers best. 2.96 was, I understand, the best gcc for IA-64 for a while. It just happened to have problems in other areas, unfortunately...
My point, though, is that you can code to real Web standards that work in IE recognizes and still do neat things. What specifically are the elements you're talking about?
Gee. I really wish only 0.0001% of sites used ECMAScript:3
I take by your comment that you run your own website. If you code it right and stick to standard HTML 4.01 Transitional (Strict is cleaner IMO, but hey) and CSS 1, everybody will see your site the same way. Literally everybody. I think it's safe to say that even IE 6 has gotten to that point. If you can't code to W3C standards, I'll be less sympathetic for your position.
Frankly, Gecko has a lot of code built in to it to do just what you say. You know how most people don't care about correct HTML these days? Gecko has a rendering mode it hits a lot that's designed to deal exactly with that. So have you had problems with recent versions of Mozilla? (Not Netscape - Mozilla. They're similar, but different.) What are these problems? I'm curious.
Seriously. Code to standards. IE is not a standard just because Microsoft wants it to be. IE understands CSS. Stick to correct HTML and CSS 1 and I guarantee everyone will see your page the way you want to. If they don't, don't whine because you won't write correct markup.
I've been able to find a page: the little CSS dropdown shopping cart at HMV. Curiously, Gecko and Safari render this properly, but Konq doesn't... adding things works about half the time, stuff like that. Generally, though, I'm fairly pleased with Konqueror, now once the tab interface matures...:3
Remember, that's just the changelog from -rc4 to -rc5. There's more changes between.23 and.24, unfortunately Alan doesn't seem to keep as accurate changelogs in 2.2.
I am guessing that 2.6 will come out in 2004, be stable and available in 2005 and many servers will only start migrating to it in 2006...
Theory has it that 2.6 will come out around the end of June 2003. Well, this is theory, of course, but there aren't really many show-stoppers left in 2.5 nowadays. I think "by the end of the year" might be a reasonable release expectation for 2.6.0 barring major problems cropping up in the next couple of months. If this is the case, I bet it would be stable enough to use in production by mid-2004.
Further, people running big machines (by "big" I mean "NR_CPUS >= 4") will probably want to upgrade; Linux scalability just keeps getting better with people like Bill Irwin and Martin Bligh hacking away...
The biggest issue with devfs, according to complaints on lkml, is that the code itself is ugly and doesn't comply with Documentation/coding-standards.txt, and that Richard Gooch isn't necessarily the greatest maintainer in the world. These days Adam Richter seems to be working on a "small-devfs" cleanup to remove cruft and such. As for stability, who knows. I don't see very much devfs-specific changes going into the kernel.
There aren't that manny artists producing thematic albums, instead of "compilations of 3-5 minute songs we just wrote."
But one of them is one of my favorite metal-ish bands! Bunch of Italian guys headed by Lucca Turilli called Rhapsody. "Hollywood Symphonic Metal" is what they call it. They tell the story of the fall and narrow salvation of this one civilization over the course of two or three of their CDs.
And... you've got to love the lyrics. o/~ Face me evil bastards, rapers of my wisdom, glory, pride, and bloodshed! Seriously, it's pretty neat. Malice Mizer (one of Japan's most famous visual kei bands) also had pretty thematic albums that were well done. Too bad we don't have more artists pushing the edge like this.
E.g. Stallman insists that man pages are obsolete and refuses to support them, which is incredibly wrongheaded.
Ironically, the last person I heard complain about that was a Debian developer. I seem to recall he also said that Debian policy is against this and in favor of having man pages for everything anyway. There's probably a happy medium... but I definitely agree (and so does Debian?) that a goal of the extirpation of man pages is silly.
Personally, I don't care how "pure" my "Unix" is either. It works the same way, and I like that:3
Linux spends almost no money in R&D and Sun spends like 2 billion. Stop ripping their shit off and come up with your own stuff or Unix will die.
Sorry, but... what the fuck? So free Unix-alikes are "ripping shit off" of Sun, now? I guess the fact that real talent contributes code to Linux doesn't excuse the fact that Linux is based around the "everything is a file" concept. So reading information in public Usenix papers is ripping off of Sun? Please. For example, the anticipatory i/o scheduler seems to be based on information that's been freely published. Not information hidden away under proprietary NDAs. Futexes and the O(1) scheduler are other examples of information that wasn't ripped-off shit. (I could be wrong, but I'm pretty sure about this.)
If Sun is spending two billion dollars in R&D and the linux people aren't, why hasn't Solaris managed to totally blow Linux out of the water? Oh wait, it does. On big (as in many processors) systems. It doesn't do as well on commodity hardware, but everybody knows Linux just doesn't scale well to 64-node machines these days. (People are working on it, but we're not there yet.) Even in the days of secure, portable, light reimplementations with wide hardware support, propietary Unix still has its niches. Besides, part of the appeal of Sun is a "total-package" deal - kind of like Apple.
Look, I appreciate that you might actually care about this, but if you don't give examples of what you're talking about you're going to look like you're talking out of your ass. Even on Slashdot.:3
I've noticed nobody ever says anything interesting about GNUStep. I'm slowly starting to look into this myself... source-level compatibilty with Mac OS X. Seems potentially interesting.
In my experience, Konqueror's rendering times are faster than Galeon's. Other things aren't as fast. (And yes, somebody needs to fix Konqueror's tabs. Fast.:3 ) Galeon crashes a fair amount on my Debian system. GTK programs have a tendency to do that. Haven't run KDE stuff on that system much, so I can't compare.
I've only ever had one problem with advanced CSS or some such in Konqueror. (The shopping cart at HMV which is some drop-down thing.)
Konqueror looks like IE? What version of IE are you talking about? The Longhorn version that doesn't exist?
MAME? You mean that arcade emulator thing? That's not exactly a major, universal application ~_^
No offense, but I wouldn't make any judgements based on the preferences of the Mplayer maintainers, either.
XMMS will be disappearing for me as soon as I can replace it. Why? mp3 support at the moment is... weird. The mpg123 plugin opens *almost* all my files, and gives horrible artifacts when it encounters stream errors. The libmad plugin handles stream errors very well, but some files crash it and it won't recognize all of them. (No matter how much I patch it, this hasn't gone away.) Since I haven't been able to get source material for all the stuff I have mp3s for yet, I can't just make Vorbis files out of these... even if I did, there would be some artifacts (though MUCH less noticeable with libmad).
Also, I don't like XMMS' refusal to pick up on toolkit appearance. That's probably also a reason why "KDE try (sic) to make a KDE tool for everything"... so that a user's apps all look alike, which historically has been one of the biggest complaints about Unix GUIs.
No, this is different. Ingo's O(1) scheduler is Linux's process scheduler, what determines when different processes get to use the CPU. The anticipatory scheduler is Nick Piggins' baby; it's an I/O scheduler, which decides when processes get to read to and write from devices.
Part of the idea is: prefer reads to writes, and assume that if you read once, more reads will follow (usually true), so wait for the next read(s) to come in and service them all at once; often, subsequent reads are to contiguous locations on disk, so if you just service them all at once you don't have to seek back if you service one, then go do something else, then get the next read.
The University of Minnesota also does this; you have to register MAC addresses under your X.500 account, and you're given up to 6. (That's just about all I need... NICs in three computers, LAN connection on my Linksys, and 802.11b card in my laptop.)
The DHCP servers only give out IPs to MAC addresses that are registered thus. Also, you have to authenticate with your X.500 account to get an IP from the campus wireless service. This seems so obvious to me I'm surprised more people don't do it ^^;
(Also, for those who read the article, the guy from UofM that says that "we don't allow clients to act as servers"... this basically means they block port 80 incoming traffic. Nothing more. Although the service agreement for res hall networking does say that you're not allowed to serve stuff.)
It's not so much the interface I don't like as much as the fact that there's no option for it to use my GTK themes. (Well, that and the XMMS developers have said that it'll always be tied only to GTK1, which occasionally gives me grief with non-ASCII characters in filenames and metadata.)
I also have to choose between an mp3 input plugin that has bad artifacts on stream errors (mpg123), or one that crashes consistently when it loads certain files ^^; And I could bitch a little bit about its source code layout, but, eh. Stylistic differences.
Slashdot Mirror
Err. I hate to say it - no, really, I do, but when I talk to people that don't run Linux, they've either heard of Red Hat, Mandrake, or SuSE.
... and wait for it to compile. The article "for people new to FreeBSD and Unix" on freebsd.org seems to verify this. And of course, FreeBSD uses a text-mode installer. Based just on *that*, FreeBSD wouldn't fit her criteria. (Although I did like it when I used it.)
Also, this may be just because I don't seek out GUI interfaces to stuff, but when I ran FreeBSD 4.5, to install software you had to go to a CLI, type in the install command
My point is, her technique was probably closest to what someone clueless, yet still brave enough to go try it, would actually do, which is what she claimed to be doing. So, criticizing her for "let's see, not enough variety" just doesn't seem right. I don't know how many people *about* to start out with Linux would have even heard of NetBSD before.
Bottom line: never upgrade kernels immediately on ANY production machine. (Never do it unless you absolutely have to, in fact.) Case in point: 2.4.18 through 2.4.20 released with bugs in ext3, notably the journal=data bug.
Preempt isn't such a big thing anymore compared to the massive scheduler and i/o scheduler improvements that have gone into 2.5.
Documentation for 2.5 changes is here courtesy Dave Jones.
No. IDE is in a rather not pretty state, according to Mr. Cox. (Although it looks like we may get IDE taskfile by 2.6 now.)
Processing an icon takes another level of brain processing, another level of indirection.
You've never played DDR, or maybe even video games, have you? When you're familiar with it, processing an image is pretty much instantaneous, not that it's terribly slow in the first place for simple stuff like this.
Well, now you can ask her :3
Hmph. It's pretty obvious why this is posted AC, I know I wouldn't be proud of this :3
Do you realize that you just pointed out that abiword, a word processor, is almost as large as the entirety of koffice?
Do you realize that, as a part of KDE, koffice takes advantage of a lot of backend that doesn't exist for, say, Abiword or OpenOffice.org?
Do you realize that there are actually developers on the GNOME project that have no relation to Sun whatsoever?
By the way, constant use of "Slowlaris" is pretty much as bad as "MS$" or "Winblows". and when I see "M$" ...
This provides a segway into using CVS for controlling revision of large software projects.
I think our friend Al Viro would have something to say about the auditing level of the Linux kernel. And if we're talking about drivers/ in particular, it would probably involve the words "obfuscated", "brain dead", "steaming pile of shit", "warped beyond all belief" ... :)
Linux code gets a fair amount of review. But once it's there, there really isn't any auditing at all.
It's only in English because BitKeeper isn't involved -_^
Interesting choice - apparently GCC 3.3 includes a lot of work SuSE have contributed. Will this be as controversial as Redhat's compiler choice of 2.96 a while back?
Probably not. GCC 3.3 will be an actual official FSF release. By the time 12 April comes around I doubt SuSE's gcc lags behind the official FSF release much. Remember, 2.96 and 2.97 were basically continuations of the 2.9x branch, after the FSF had basically stopped working on it and started working on GCC 3.x.
Thing is, gcc 3.x broke things. Also, Red Hat had a collection of IA-64 improvements for gcc that may not have made it into mainline yet. So, they made the unofficial releases because they felt that's what served their customers best. 2.96 was, I understand, the best gcc for IA-64 for a while. It just happened to have problems in other areas, unfortunately ...
My point, though, is that you can code to real Web standards that work in IE recognizes and still do neat things. What specifically are the elements you're talking about?
Gee. I really wish only 0.0001% of sites used ECMAScript :3
I take by your comment that you run your own website. If you code it right and stick to standard HTML 4.01 Transitional (Strict is cleaner IMO, but hey) and CSS 1, everybody will see your site the same way. Literally everybody. I think it's safe to say that even IE 6 has gotten to that point. If you can't code to W3C standards, I'll be less sympathetic for your position.
Frankly, Gecko has a lot of code built in to it to do just what you say. You know how most people don't care about correct HTML these days? Gecko has a rendering mode it hits a lot that's designed to deal exactly with that. So have you had problems with recent versions of Mozilla? (Not Netscape - Mozilla. They're similar, but different.) What are these problems? I'm curious.
Seriously. Code to standards. IE is not a standard just because Microsoft wants it to be. IE understands CSS. Stick to correct HTML and CSS 1 and I guarantee everyone will see your page the way you want to. If they don't, don't whine because you won't write correct markup.
I've been able to find a page: the little CSS dropdown shopping cart at HMV. Curiously, Gecko and Safari render this properly, but Konq doesn't ... adding things works about half the time, stuff like that. Generally, though, I'm fairly pleased with Konqueror, now once the tab interface matures ... :3
Remember, that's just the changelog from -rc4 to -rc5. There's more changes between .23 and .24, unfortunately Alan doesn't seem to keep as accurate changelogs in 2.2.
I am guessing that 2.6 will come out in 2004, be stable and available in 2005 and many servers will only start migrating to it in 2006 ...
Theory has it that 2.6 will come out around the end of June 2003. Well, this is theory, of course, but there aren't really many show-stoppers left in 2.5 nowadays. I think "by the end of the year" might be a reasonable release expectation for 2.6.0 barring major problems cropping up in the next couple of months. If this is the case, I bet it would be stable enough to use in production by mid-2004.
Further, people running big machines (by "big" I mean "NR_CPUS >= 4") will probably want to upgrade; Linux scalability just keeps getting better with people like Bill Irwin and Martin Bligh hacking away ...
The biggest issue with devfs, according to complaints on lkml, is that the code itself is ugly and doesn't comply with Documentation/coding-standards.txt, and that Richard Gooch isn't necessarily the greatest maintainer in the world. These days Adam Richter seems to be working on a "small-devfs" cleanup to remove cruft and such. As for stability, who knows. I don't see very much devfs-specific changes going into the kernel.
There aren't that manny artists producing thematic albums, instead of "compilations of 3-5 minute songs we just wrote."
But one of them is one of my favorite metal-ish bands! Bunch of Italian guys headed by Lucca Turilli called Rhapsody. "Hollywood Symphonic Metal" is what they call it. They tell the story of the fall and narrow salvation of this one civilization over the course of two or three of their CDs.
And ... you've got to love the lyrics. o/~ Face me evil bastards, rapers of my wisdom, glory, pride, and bloodshed! Seriously, it's pretty neat. Malice Mizer (one of Japan's most famous visual kei bands) also had pretty thematic albums that were well done. Too bad we don't have more artists pushing the edge like this.
E.g. Stallman insists that man pages are obsolete and refuses to support them, which is incredibly wrongheaded.
Ironically, the last person I heard complain about that was a Debian developer. I seem to recall he also said that Debian policy is against this and in favor of having man pages for everything anyway. There's probably a happy medium ... but I definitely agree (and so does Debian?) that a goal of the extirpation of man pages is silly.
Personally, I don't care how "pure" my "Unix" is either. It works the same way, and I like that :3
Linux spends almost no money in R&D and Sun spends like 2 billion. Stop ripping their shit off and come up with your own stuff or Unix will die.
Sorry, but ... what the fuck? So free Unix-alikes are "ripping shit off" of Sun, now? I guess the fact that real talent contributes code to Linux doesn't excuse the fact that Linux is based around the "everything is a file" concept. So reading information in public Usenix papers is ripping off of Sun? Please. For example, the anticipatory i/o scheduler seems to be based on information that's been freely published. Not information hidden away under proprietary NDAs. Futexes and the O(1) scheduler are other examples of information that wasn't ripped-off shit. (I could be wrong, but I'm pretty sure about this.)
If Sun is spending two billion dollars in R&D and the linux people aren't, why hasn't Solaris managed to totally blow Linux out of the water? Oh wait, it does. On big (as in many processors) systems. It doesn't do as well on commodity hardware, but everybody knows Linux just doesn't scale well to 64-node machines these days. (People are working on it, but we're not there yet.) Even in the days of secure, portable, light reimplementations with wide hardware support, propietary Unix still has its niches. Besides, part of the appeal of Sun is a "total-package" deal - kind of like Apple.
Look, I appreciate that you might actually care about this, but if you don't give examples of what you're talking about you're going to look like you're talking out of your ass. Even on Slashdot. :3
I love Window Maker ^_^
... source-level compatibilty with Mac OS X. Seems potentially interesting.
I've noticed nobody ever says anything interesting about GNUStep. I'm slowly starting to look into this myself
In my experience, Konqueror's rendering times are faster than Galeon's. Other things aren't as fast. (And yes, somebody needs to fix Konqueror's tabs. Fast. :3 ) Galeon crashes a fair amount on my Debian system. GTK programs have a tendency to do that. Haven't run KDE stuff on that system much, so I can't compare.
... weird. The mpg123 plugin opens *almost* all my files, and gives horrible artifacts when it encounters stream errors. The libmad plugin handles stream errors very well, but some files crash it and it won't recognize all of them. (No matter how much I patch it, this hasn't gone away.) Since I haven't been able to get source material for all the stuff I have mp3s for yet, I can't just make Vorbis files out of these ... even if I did, there would be some artifacts (though MUCH less noticeable with libmad).
... so that a user's apps all look alike, which historically has been one of the biggest complaints about Unix GUIs.
I've only ever had one problem with advanced CSS or some such in Konqueror. (The shopping cart at HMV which is some drop-down thing.)
Konqueror looks like IE? What version of IE are you talking about? The Longhorn version that doesn't exist?
MAME? You mean that arcade emulator thing? That's not exactly a major, universal application ~_^
No offense, but I wouldn't make any judgements based on the preferences of the Mplayer maintainers, either.
XMMS will be disappearing for me as soon as I can replace it. Why? mp3 support at the moment is
Also, I don't like XMMS' refusal to pick up on toolkit appearance. That's probably also a reason why "KDE try (sic) to make a KDE tool for everything"
No, this is different. Ingo's O(1) scheduler is Linux's process scheduler, what determines when different processes get to use the CPU. The anticipatory scheduler is Nick Piggins' baby; it's an I/O scheduler, which decides when processes get to read to and write from devices.
Part of the idea is: prefer reads to writes, and assume that if you read once, more reads will follow (usually true), so wait for the next read(s) to come in and service them all at once; often, subsequent reads are to contiguous locations on disk, so if you just service them all at once you don't have to seek back if you service one, then go do something else, then get the next read.
The University of Minnesota also does this; you have to register MAC addresses under your X.500 account, and you're given up to 6. (That's just about all I need ... NICs in three computers, LAN connection on my Linksys, and 802.11b card in my laptop.)
... this basically means they block port 80 incoming traffic. Nothing more. Although the service agreement for res hall networking does say that you're not allowed to serve stuff.)
The DHCP servers only give out IPs to MAC addresses that are registered thus. Also, you have to authenticate with your X.500 account to get an IP from the campus wireless service. This seems so obvious to me I'm surprised more people don't do it ^^;
(Also, for those who read the article, the guy from UofM that says that "we don't allow clients to act as servers"
It's not so much the interface I don't like as much as the fact that there's no option for it to use my GTK themes. (Well, that and the XMMS developers have said that it'll always be tied only to GTK1, which occasionally gives me grief with non-ASCII characters in filenames and metadata.)
I also have to choose between an mp3 input plugin that has bad artifacts on stream errors (mpg123), or one that crashes consistently when it loads certain files ^^; And I could bitch a little bit about its source code layout, but, eh. Stylistic differences.