...but every time someone slams AOL, they're essentially saying "Go MSN!".
AOL has been pretty benevolent so far - vastly more so than microsoft. They deserve to be treated well until they let us down in a big way. Because AOL is our greatest hope in the battle against microsoft. They can single-handedly win the browser war against microsoft, among other things.
They never miss a chance to slam solaris - which is most admins' bread and butter. And the gratuituous attacks are far from based on facts - they're extremely misleading.
Wow, I so didn't realize that you get what you pay for. In that case, I think I'll get rid of all my linux machines and commission a port of windows xp to the mainframe and 3270 terminals I'll have to buy - that'd cost plenty, so it must be great stuff.
Go RMS. He goes too far sometimes, but this time he's spot-on.
Miguel... Geeze, did he sign some sort of secret deal with microsoft? It's -insane- to become dependent on them. Look at the huge trail of partners microsoft has destroyed ("innovated").
I like gnome. I've invested time in learning gnome programming. But this has got me having second thoughts about maybe switching to KDE. I believe in gnome because it's more open. A gnome that requires.net isn't.
If microsoft is onto something with.net, there's no good reason why we can't clone the ideas and ignore their API's - As long as they haven't patented something, in which case we shouldn't be using it anyway.
Does anyone know the most effective places to send letters to make sure gnome doesn't become dependent on.net? Like Miguel himself, the Gnome Foundation (I don't have addresses for either), and maybe anything else that might be a help?
In the universities I went to, SDSM&T and U of Cinci, if you clept out of something, that would allow you to skip a particular class, but it didn't allow you to reduce the total number of classes you had to sit through and pay for.
I'd expect most accredited universities, if not all, to require that you sit through the same number of classes and buy your degree if you already have the knowledge. And you can expect the classes to get mired in the details of specific examples that get away from the point, that'll prevent you from skipping class a lot, even if you already know the real subject matter pretty well.
I found this really frustrating when I went to college, but I couldn't find a way around it. I think the best you can do is to take the toughest classes you can find, fight the tedium, and study your ass off to get the best education you can instead of coasting on what you already know. You're paying a lot for it; may as well get a lot out of it.
Nah. If we move people from something that's entirely proprietary, to something that's mostly nonproprietary, that's a big step in the right direction. You've presented another "all or nothing" argument.
If I have a choice between supporting endusers who run a microsoft os, and endusers who run lindows, which do you think I'd rather choose?
I don't see myself giving up redhat with fvwm2 for my own personal use anytime soon, but I do plan to evaluate the preview release of lindows as soon as its made available, because I would love to be able to recommend it to my boss's boss for a big on-campus usage push. He was just asking about how we should satisfy the cries for more linux support on campus the other day.
BTW, if wine isn't GPL'd, and perhaps even if it is, I'd lay odds that the lindows folks just grabbed wine and banged on it a bit to make it run a few more applications. But if it isn't GPL'd, I'd guess they won't be releasing their changes back to the community they borrowed it from.
...but it seems to me you'd be in good shape for the future if you used some good wide PVC for your existing cables, and ran some good strong string (fishing line maybe), so you can pull whatever you need in the future. Then you don't have to worry about something replacing fiber, or your connectors becoming outdated, or whatever.
When it comes time to string your fiber or whatever, you pull two things - the fiber, and a replacement string for the next time you need to upgrade. Just tie'em both to the fishing line that's already in there, and pull.
I've written quite a bit of/bin/sh code, and have found it easy to stick to a portable common subset of bourne shell and bash. Just stay away from the truly esoteric stuff, and you're fine.
1) A lot of universities have a sysadmin class or two. Take what you can.
2) Read newsgroups related to sysadmining. This is the single most important recommendation, IMO.
3) Attend sysadmin conferences
4) Programming classes are useful, but optional. You can become an advanced sysadmin in some places just by knowing how to interpret truss/strace/trace/par output. If you can also write glue code in C,/bin/sh, and python, you're in really good shape.
5) Get a box you can yank on really hard without upsetting anyone. Do your evil experiments there, not on production machines. If you want to admin linux, get a linux box. If you want to admin sun, get a sun, even if it's kind of low-end, as long as it'll run a current release of solaris.
6) I have a rather large collection of intro links here.
7) A four year degree helps, but you wouldn't necessarily have to make it computer related. One of our better admins here has a poli sci degree. With this degree, he showed he could jump through hoops, which is the most important thing a degree does for you, IMO. Then again, I went for an MS in CS, and I haven't regretted it, despite its not being all that directly applicable.
I have the same trouble on a redhat system with mutt (another mail program) if I use a too-recent version of nfs-utils. I think this was only on 6.2, IIRC. I don't recall having the problem on redhat 7.1.
1) Write something for fun when you have some time - Maybe over the winter solstice holiday. This way -you- define how good it is and what your project needs, not someone who's looking for something easy to grade. It's nice to write creatively, and not (only) some mind-crushingly mass-produced assignment.
2) Work in a higher level language when you can. Assembler's ludicrous. C isn't great either. C++ isn't much better. java is a middle ground. Really good choices include python and ML - few words per idea expressed is your goal. Go with perl if you must go along with the crowd; it's HL, but pretty ugly, like sendmail.cf.
3) Analyze your mistakes and learn from them. EG, if you're correcting your own off-by-ones regularly, inculcate a habit of thinking, every time you encounter a boundary condition, "how can this go wrong?" and try to prevent the problem before it comes up. You might even keep tally of the bugs you run across in your own code, and focus on prevention of the ones that come up the most. I know this is a bit of a pain, but it's a lot less of a pain than tracking down a runtime error, especially once you get in the habit.
IANAL, but I believe PD means it isn't under a license.
IANAL, but I believe Public Domain means any unscrupulous person or company can come along and put said software package under their own license, charge money for the software, and more importantly, steal away all rights of the original authors.
It's much better to GPL your software, or at least put it under a BSD like license, to protect your own rights, and the rights of those who use your software.
I'm always very amused when I see today's crop of movie-like games referred to as "strategy games".
If you want something heavy on strategy, minus the explosions and dudes keeling over with bullet holes, try the aeons-old board game called go. There are computer opponents for *ix which are good enough to challenge a beginner, and there are clients for go servers that'll give you access to some astonishingly good human opponents as well.
It's incredible how a game with such simple rules can have so many layers of deep strategy. It actually uses a rating system just like what you see used by students of karate, to help give people well-matched games.
Folks, GP is a really promising area, but please don't forget, it's a patented algorithm, and probably will be for some time.
I wrote to the patent holder to see if I could use GP in a GPL'd project (offered him a big chunk of my prize money if my project panned out), and after suggesting I publish my results if I get anywhere instead, he totally blew me off. The project in question was a "go/igo/weichi/baduk AI", a really challenging artificial intelligence problem that has eluded many. I'm actually more interested in giving away my code than the prize money though.
So you can probably assume that where OSS and FS are concerned, this guy isn't interested, although maybe if you're project doesn't have prize money involved, he might react differently; I don't know. I'm beginning to think lawyers tell their clients not to tip their hand either way where patents are concerned, because it seems there's a big vacuum surrounding FS/OSS and patents.:(
Actually, I've been running a beta of xemacs 21 with gtk support for months - it's worked great, too. So it's there, you probably just need to look a little harder.
You can get go client software for just about any common platform, some are in java too. So you can play on servers against other humans, and you can play against AI programs (which most people can beat easily after playing a few years because go is just too deep for a computer to play well - despite a $1 million reward for a truly good go AI).
There have been studies on the effects of go on the brain. Apparently it has developmental benefits in children similar to studying music. It supposedly also helps with dementia.
If the antitrust suit's remedy is toothless, will the Micro$oft Monitor be reactivated?
In the M$ Monitor's goodbye message, the author alluded to other news/coordination resources that were duplicating the M$ Monitor's function. What are these resources?
I miss the M$ Monitor. Of course, if I have to choose between having the Monitor back and having an effective penalty against Microsoft, I'll choose the latter.:)
Actually, it's about time a cpu vendor started doing this.
Benchmarks suck, but measuring your performance soley by MHz sucks worse.
Hopefully now benchmarking will be revived, and evolve into something more meaningful than it already is. Perhaps the recently released AIM benchmarks will play a role, who knows. nbench isn't horrible (compared to pure MHz comparisons). Forget bogomips - they might actually be worse than a pure MHz comparison.
"Strong competition" and "innovation" have not been the twin hallmarks of the software industry in the recent past. How could you even suggest such a thing.
No, "marketing muscle" and "exclusionary licensing practices" have been the twin hallmarks of the software industry in the recent past. And perhaps we should throw in "embrace, extend, extinguish".
The free market has failed where software is concerned. A mostly-free market with just enough regulation to prevent unfair competition from monopolies is an absolute requirement for a healthy software industry.
I don't think a quick reference is practical for something that is not only very complicated, but also to some extent a matter of personal preference.
For example, this quick reference puts into doubt whether "the community" (which community?!) likes to accept code under the GPL, and indicates that there is no problem accepting code under a BSD license.
IMO, it should be the other way around. GPL code is safe from an unscrupulous vendor doing what microsoft did to kerberos (use it, don't give anything back, and mess up interoperability for the people who originally wrote the code). The BSD license doesn't have such a protection, so I consider it unsafe, aka not popularly accepted.
You don't have to agree. That's the point: This is not something that can be summarized effectively in such a tiny little chart.
(CERT being so in bed with the vendors that they often take years to get the word out, if they do at all. Meanwhile the black hats have full access to the info and plenty of time to start exploiting it)
He came and gave a talk about GP at UCI, so I went to see it. He gave an interesting song and dance about the great results they've gotten with it. I have no reason to doubt that the results are really good for some problem domains.
I got kind of excited about trying GP for a go engine, one of the hardest AI problems around. It's been tried once before, but I suspect the programmers didn't have that much domain-specific knowledge to help along the project.
What Koza didn't tell us, and what I didn't find out until I'd already invested a fair amount of time into studying GP, is this:
GP itself is patented.
So you can't really use GP as a way of coming up with patents to defend free software, unless you license the patent for GP.
BTW, I wrote Koza and asked him if I could use GP in a free software project for that go engine I was drooling about. I offered to split that $1M+ prize with him if I came up with something decent. He didn't even reply.
Slashdot Mirror
...but every time someone slams AOL, they're essentially saying "Go MSN!".
AOL has been pretty benevolent so far - vastly more so than microsoft. They deserve to be treated well until they let us down in a big way. Because AOL is our greatest hope in the battle against microsoft. They can single-handedly win the browser war against microsoft, among other things.
Actually, the Nemeth book really sucks.
They never miss a chance to slam solaris - which is most admins' bread and butter. And the gratuituous attacks are far from based on facts - they're extremely misleading.
Wow, I so didn't realize that you get what you pay for. In that case, I think I'll get rid of all my linux machines and commission a port of windows xp to the mainframe and 3270 terminals I'll have to buy - that'd cost plenty, so it must be great stuff.
Go RMS. He goes too far sometimes, but this time he's spot-on.
Miguel... Geeze, did he sign some sort of secret deal with microsoft? It's -insane- to become dependent on them. Look at the huge trail of partners microsoft has destroyed ("innovated").
I like gnome. I've invested time in learning gnome programming. But this has got me having second thoughts about maybe switching to KDE. I believe in gnome because it's more open. A gnome that requires
If microsoft is onto something with
Does anyone know the most effective places to send letters to make sure gnome doesn't become dependent on
In the universities I went to, SDSM&T and U of Cinci, if you clept out of something, that would allow you to skip a particular class, but it didn't allow you to reduce the total number of classes you had to sit through and pay for.
I'd expect most accredited universities, if not all, to require that you sit through the same number of classes and buy your degree if you already have the knowledge. And you can expect the classes to get mired in the details of specific examples that get away from the point, that'll prevent you from skipping class a lot, even if you already know the real subject matter pretty well.
I found this really frustrating when I went to college, but I couldn't find a way around it. I think the best you can do is to take the toughest classes you can find, fight the tedium, and study your ass off to get the best education you can instead of coasting on what you already know. You're paying a lot for it; may as well get a lot out of it.
Yes. And I wouldn't want to be the company in the position Lindows is taking.
However, I love basking in the glow of the lawsuit that can be expected to arise when Microsoft tries this tactic.
Nah. If we move people from something that's entirely proprietary, to something that's mostly nonproprietary, that's a big step in the right direction. You've presented another "all or nothing" argument.
What is it with all-or-nothing people?
If I have a choice between supporting endusers who run a microsoft os, and endusers who run lindows, which do you think I'd rather choose?
I don't see myself giving up redhat with fvwm2 for my own personal use anytime soon, but I do plan to evaluate the preview release of lindows as soon as its made available, because I would love to be able to recommend it to my boss's boss for a big on-campus usage push. He was just asking about how we should satisfy the cries for more linux support on campus the other day.
BTW, if wine isn't GPL'd, and perhaps even if it is, I'd lay odds that the lindows folks just grabbed wine and banged on it a bit to make it run a few more applications. But if it isn't GPL'd, I'd guess they won't be releasing their changes back to the community they borrowed it from.
...but it seems to me you'd be in good shape for the future if you used some good wide PVC for your existing cables, and ran some good strong string (fishing line maybe), so you can pull whatever you need in the future. Then you don't have to worry about something replacing fiber, or your connectors becoming outdated, or whatever.
When it comes time to string your fiber or whatever, you pull two things - the fiber, and a replacement string for the next time you need to upgrade. Just tie'em both to the fishing line that's already in there, and pull.
I've written quite a bit of
2) Read newsgroups related to sysadmining. This is the single most important recommendation, IMO.
3) Attend sysadmin conferences
4) Programming classes are useful, but optional. You can become an advanced sysadmin in some places just by knowing how to interpret truss/strace/trace/par output. If you can also write glue code in C, /bin/sh, and python, you're in really good shape.
5) Get a box you can yank on really hard without upsetting anyone. Do your evil experiments there, not on production machines. If you want to admin linux, get a linux box. If you want to admin sun, get a sun, even if it's kind of low-end, as long as it'll run a current release of solaris.
6) I have a rather large collection of intro links here.
7) A four year degree helps, but you wouldn't necessarily have to make it computer related. One of our better admins here has a poli sci degree. With this degree, he showed he could jump through hoops, which is the most important thing a degree does for you, IMO. Then again, I went for an MS in CS, and I haven't regretted it, despite its not being all that directly applicable.
Doesn't sound like an evolution bug to me.
Sounds like an NFS problem.
I have the same trouble on a redhat system with mutt (another mail program) if I use a too-recent version of nfs-utils. I think this was only on 6.2, IIRC. I don't recall having the problem on redhat 7.1.
1) Write something for fun when you have some time - Maybe over the winter solstice holiday. This way -you- define how good it is and what your project needs, not someone who's looking for something easy to grade. It's nice to write creatively, and not (only) some mind-crushingly mass-produced assignment.
2) Work in a higher level language when you can. Assembler's ludicrous. C isn't great either. C++ isn't much better. java is a middle ground. Really good choices include python and ML - few words per idea expressed is your goal. Go with perl if you must go along with the crowd; it's HL, but pretty ugly, like sendmail.cf.
3) Analyze your mistakes and learn from them. EG, if you're correcting your own off-by-ones regularly, inculcate a habit of thinking, every time you encounter a boundary condition, "how can this go wrong?" and try to prevent the problem before it comes up. You might even keep tally of the bugs you run across in your own code, and focus on prevention of the ones that come up the most. I know this is a bit of a pain, but it's a lot less of a pain than tracking down a runtime error, especially once you get in the habit.
IANAL, but I believe PD means it isn't under a license.
IANAL, but I believe Public Domain means any unscrupulous person or company can come along and put said software package under their own license, charge money for the software, and more importantly, steal away all rights of the original authors.
It's much better to GPL your software, or at least put it under a BSD like license, to protect your own rights, and the rights of those who use your software.
If you want something heavy on strategy, minus the explosions and dudes keeling over with bullet holes, try the aeons-old board game called go. There are computer opponents for *ix which are good enough to challenge a beginner, and there are clients for go servers that'll give you access to some astonishingly good human opponents as well.
It's incredible how a game with such simple rules can have so many layers of deep strategy. It actually uses a rating system just like what you see used by students of karate, to help give people well-matched games.
I wrote to the patent holder to see if I could use GP in a GPL'd project (offered him a big chunk of my prize money if my project panned out), and after suggesting I publish my results if I get anywhere instead, he totally blew me off. The project in question was a "go/igo/weichi/baduk AI", a really challenging artificial intelligence problem that has eluded many. I'm actually more interested in giving away my code than the prize money though.
So you can probably assume that where OSS and FS are concerned, this guy isn't interested, although maybe if you're project doesn't have prize money involved, he might react differently; I don't know. I'm beginning to think lawyers tell their clients not to tip their hand either way where patents are concerned, because it seems there's a big vacuum surrounding FS/OSS and patents. :(
I believe the patent in question is 4,935,877.
Actually, I've been running a beta of xemacs 21 with gtk support for months - it's worked great, too. So it's there, you probably just need to look a little harder.
No mailman?
'seems to be the big one these days. Maybe it wasn't, back when the book was written.
My personal pick is Go. It's Soooo deep.
You can get go client software for just about any common platform, some are in java too. So you can play on servers against other humans, and you can play against AI programs (which most people can beat easily after playing a few years because go is just too deep for a computer to play well - despite a $1 million reward for a truly good go AI).
There have been studies on the effects of go on the brain. Apparently it has developmental benefits in children similar to studying music. It supposedly also helps with dementia.
In the M$ Monitor's goodbye message, the author alluded to other news/coordination resources that were duplicating the M$ Monitor's function. What are these resources?
I miss the M$ Monitor. Of course, if I have to choose between having the Monitor back and having an effective penalty against Microsoft, I'll choose the latter. :)
Actually, it's about time a cpu vendor started doing this.
Benchmarks suck, but measuring your performance soley by MHz sucks worse.
Hopefully now benchmarking will be revived, and evolve into something more meaningful than it already is. Perhaps the recently released AIM benchmarks will play a role, who knows. nbench isn't horrible (compared to pure MHz comparisons). Forget bogomips - they might actually be worse than a pure MHz comparison.
"Strong competition" and "innovation" have not been the twin hallmarks of the software industry in the recent past. How could you even suggest such a thing.
No, "marketing muscle" and "exclusionary licensing practices" have been the twin hallmarks of the software industry in the recent past. And perhaps we should throw in "embrace, extend, extinguish".
The free market has failed where software is concerned. A mostly-free market with just enough regulation to prevent unfair competition from monopolies is an absolute requirement for a healthy software industry.
I don't think a quick reference is practical for something that is not only very complicated, but also to some extent a matter of personal preference.
For example, this quick reference puts into doubt whether "the community" (which community?!) likes to accept code under the GPL, and indicates that there is no problem accepting code under a BSD license.
IMO, it should be the other way around. GPL code is safe from an unscrupulous vendor doing what microsoft did to kerberos (use it, don't give anything back, and mess up interoperability for the people who originally wrote the code). The BSD license doesn't have such a protection, so I consider it unsafe, aka not popularly accepted.
You don't have to agree. That's the point: This is not something that can be summarized effectively in such a tiny little chart.
Sounds like the path to CERT to me.
(CERT being so in bed with the vendors that they often take years to get the word out, if they do at all. Meanwhile the black hats have full access to the info and plenty of time to start exploiting it)
Personally, I'm a bit annoyed with Koza.
He came and gave a talk about GP at UCI, so I went to see it. He gave an interesting song and dance about the great results they've gotten with it. I have no reason to doubt that the results are really good for some problem domains.
I got kind of excited about trying GP for a go engine, one of the hardest AI problems around. It's been tried once before, but I suspect the programmers didn't have that much domain-specific knowledge to help along the project.
What Koza didn't tell us, and what I didn't find out until I'd already invested a fair amount of time into studying GP, is this:
GP itself is patented.
So you can't really use GP as a way of coming up with patents to defend free software, unless you license the patent for GP.
BTW, I wrote Koza and asked him if I could use GP in a free software project for that go engine I was drooling about. I offered to split that $1M+ prize with him if I came up with something decent. He didn't even reply.