Buy 1 lab worth of AMD machines, and see how it works out. You've probably got a dozen different models of PCs out there anyway; one more flavor isn't going to hurt you too much.
If things go badly, much better to find out with 40 machines then 400.
>The way my company gets around this is when I
>have training, the company pays for it. But if I
>leave within 6 months of having the training I
>pay a portion of the cost: all the price if I
>leave immediately, 5/6 if I leave 1 month after
>the training, 4/6 after 2 months, etc. This way
>they get value for their money, and I get to
>leave if I want, but I have to pay my dues!
If someone else really wants to hire you away, they'll gladly pay that bill for your training.
That sort of thing is more of a speedbump then a barrier to a new job.
Why bother scanning email for porn when you could just scan the cache of your web proxy every few hours? In my experience there's a lot more porn coming in over port 80 then over the email system.
I have no idea about the content of your site, but I see a lot of sites where everything is dynamicly generated, for no obvious reason.
Maybe you could save a few thousand hits on your database by making some of the pages static, or just regenerating them with a perl script once an hour or something.
>Use their example, say, a spare part for your
>dishwasher. An 'L' pipe connector for instance.
>Your L connector break? Go to www.maytag.com,
>download the instructions for your 3D printer to
>construct a new part! Manufacturer's construction
>costs are eliminated!
Yeah, but Maytag will never go for it because spare parts are a revenue source for them. I'm sure they make a ton of money selling "L" pipe connectors at 200% markup.
>How does this program tell the difference between
>an intruder modifying files using a real/spoofed
>login and a normal user modifying his own files
>that he should be modifying? Or is this program
>not designed to catch that?
Not familiar with this particular s/w, but with this sort of thing you can generally pick and choose which files/directories to watch. You're not going to bother checksumming/home because you don't really care.
And if you're the admin, you're going to remember what you did. If you add a new HD or something and get an alert the next day saying that/etc/fstab has been modified, it (hopefully!) won't be a surprise. It's the file changes you can't account for that you're supposed to worry about.
>The only way to do it is to have a group of
>people who are familiar with the ports systems of
>the respective bsd systems, BUT that doesnt hold
>any strong arogance towards a specific BSD. The
>project must be done independantly and not tied
>to a specific BSD.
I saw something here a few months ago about some Debian folks playing with the idea of a ports infrastructure for Debian... Might as well put them into the Grand Unified Ports Project too.
That's the whole thing, finding a bunch of people with a) the technical know-how, b) the time, and c) the diplomatic skills to pull this off.
Testing would be a cast-iron bitch: I can only imagine what could go wrong on all of those OSs and platforms.
I hate to sound cynical, but this is going to be one of those really good ideas that comes up every 6 months or so and dies down again because nobody is ambitious (crazy?) enough to tackle it.
>If the problem still looks bad when I get back to >it, then the next thing I do is ask a colleague >for advice. In situations where there isn't >another programmer around, I simply describe the >problem to any willing listener. Often the simple >act of talking about a problem elucidates the >solution. Now this is going to sound silly, but >I've even talked to my cat when working at home >and sometimes had that "eureka" moment.
I heard of a university help desk who used a teddy bear as their first line of support; anyone with a programming problem had to explain it to the bear before a human would deal with them. Apparantly the bear solved lots of problems, too.
No, I don't remember where I heard this story, maybe it's just an urban legend.
>Know where the UPS keys and manuals and call >sheets are. Laminate them and tape them to the >wall, right by the flashlights (more that one) >that you check once a week.
>Question: What are the options for a small-to- >medium server room? (about 15-30 middling sized >computers) Right now we have a single largish >UPS, which says it's running at 40-60% of >capacity. (depending on if I have the staging DB >powered up) As we gain 2 more of those systems, >and several other power-hungry systems, I know >that our current UPS will not provide more than >a minute or two of power for everything. We are >not even monitoring it at the moment (we just >hope that the power doesn't go out).
UPSen are rated in terms of KVA, and most of your equipment will probably have some info on the power supply stating it's power requirements.
Or you can bring in your friendly neighbourhood UPS salescritter, they have lots of practice making these calculations.
Monitors are pigs, if you can consolidate a few monitors with a switchbox then you'll save surprising amounts of power.
MS Poxy supports the usual CERN http/ftp proxy stuff.
MS Poxy also has a proprietary proxy protocol that only works with their (vile) Windows client.
MS Poxy also supports Socks 4, but not Socks 5. Ask your admin guys about the settings, and try and find a Socks 4 client for whatever you're doing. I've gotten Mirc to work over Socks 4, so it can be done...
The proxy documentation is probably hidden somewhere on the MS web site, or if your admin guys are co-operative it's also on the CD.
>Yes, I could write a bash script or perl script >that deletes files. Guess what, not everyone uses >bash and has perl on their unix system, and if >they did, it would only delete their user files, >and NO system files would be affected.
Well, you could find a *nix common denominator, i.e. pretty much everything has a/bin/sh and/bin/rm don't they?
My 2 cents: we shouldn't get too complacent about this type of thing. Another nasty *nix worm/virus of some sort is probably inevitable. The more that people celebrate the problems that MS is having right now, the harder we're all going to have our noses rubbed in it when the next Morris worm comes around.
I'll gladly criticize problems/w Outlook's design, but I'll never say "that could never happen here" because it's too much like tempting fate.
>What is the difference between a version (BSD) >and a distro (Linux)?
The various Linux distros all have the same Linux kernel (maybe different versions, but it's all the same kernel, more or less).
Linux distros differ mostly in userland stuff, i.e. how the directories are laid out, init scripts, what utilities are bundled, packaging infrastructure, etc.
The various BSDs have different kernels. They all started from the same codebase, but have diverged significantly since.
Of course, the various BSD userland stuff has some differences as well.
Re:Open Sources Has a pretty good BSD Chapter
on
The Roots Of BSD
·
· Score: 2
>The Chapter Twenty Years of Berkley Unix: From >AT&T Owned to Freely Redistribuatble by Mcusick & >Co. does a very good job at describing the >movement and development from the AT&T based >Berkley Unix to the Free versions that we have >today.
>I think, though, that the reason Linux can have >such a frequent release schedule is that it >doesn't have a marketing behemoth behind it. It's >target audience is largely comprised of >technophiles who can pick up a new release, apply >it correctly and troubleshoot the possible >interoperability problems that arise with >existing installed software. The Windows family >is largely aimed at users who can't.
>The vehicle for distribution is different, too. >The fact that Joe Windows User buys a box of >software means that he does not want to have >to buy another box of software in six months when >the "new release" (read: service pack) comes out. >Linux is mostly downloaded. Download the product. >Download the updates. Because Windows is >packaged, vast amounts of money would be wasted >on repackaging for every service pack while older >"releases" sat on the shelf never to be bought. >(Example: do stores stock Windows95 next to >Windows98 boxes? If they did, would anyone buy >95?)
You're right of course, it's a big cost for the manufacturer and distribution channel when those boxes sit on the shelf and never sell. Red Hat etc. get burned by having surplus inventory as well, I still see some RH 5.2 boxes sitting around the local computer store.
One of the great things about FreeBSD is that there's a new release every 6 months. Most of the time it's just new drivers and teeny bug fixes, i.e. there was nothing too dramatic between 3.3 and 3.4.
But these new releases give me a bootable install CD that includes ALL of the latest h/w support, which greatly simplifies installation. No more piddling around adding SCSI drivers half way thru an NT install, or trying to build a custom install diskette, etc. etc.
Of course FreeBSD doesn't have any retail market to speak of, so a new rev doesn't cost them anything.
>These computers can be useful, providing nice >email access and such, but are they really that >good in the long run? Is this where computers are >heading? If not, why are we arranging for our >children to learn on these things, and not on >real computers, which will undoubtedly have more > challenging interfaces,
I don't think history is on your side here. The average PC user in 1992 knew how to move files around at a DOS prompt, was capable of formatting their own diskettes, etc etc.
This is certainly no longer the case. People are working around the clock to make interfaces as idiot-friendly as possible. (I don't think they are succeeding, but they sure are trying.)
>will be more capable, and will also be more >complex. As a network administrator, I see >users who cannot understand the simple (in my >mind) concept of a directory tree structure. >Shouldn't we be focusing on teaching our kids >how to use actual computers? Unless, this is >where computers are heading. I like to think not, >however the current trend appears to be >simplification of computers, rather than >education of users. I think those two are both >needed, but they should meet in the middle >somewhere. Whatever. Random mumblings from a >lunatic, take them as you will.
I agree with you. If you can't understand what a directory structure is, then go use a fscking typewriter or something. The industry, and the average consumer, have different ideas, I guess.
This little puppy would have an advantage for schools, in that it's impossible to screw up the software configuration. A few labs full of tamper-proof machines would be worth hundreds of man hours a year.
>Anyone thinking this is happening very close to >april 1? I just cannot fathom the ultima series >without its original designer.
They (Origin) have been saying that Ultima IX was the last Ultima.
I'm shedding no tears, Ultima has been going steadily downhill. Those guys clearly had no intention of making another U5 or U7; I don't know if that's Garriot's fault or the marketroids'.
Hopefully a fresh start will give Garriot a chance to recapture the magic of those first few games. I admire the guy, he's clearly brilliant, but those last 2 "Ultima" games don't deserve the name.
>Hmm. Much though I dislike this tactic in >Microsoft, it is probably worth pointing out that >Novell, M$'s biggest competitor in the network >server market, are doing almost exactly the same >thing to their CNEs - "upgrade to a Netware 5 >CNE, or lose it"
Sort of, not really.
NW5 has been out since Fall 98, and you have until Aug 2000 to recertify. This requires 1 test for 4.x CNEs.
W2K has been out since Feb 2000, and you have until Dec 2000 to recertify. This requires 2 or 3 tests IIRC.
>Someone stole ALL the mouseballs from one of the >computer labs.
Someone did that at my Uni a few years back.
So the Computer Services guys glued ALL of the mice shut. In ALL of the labs in that building. For all I know, the Uni is still glueing it's new mice shut.
So there are hundreds of mice that can't be cleaned and are slowly getting gummed up and unusable... All because of some asshole prank.
>Furthurmore, someone has everyone believing that >all students are computer virus carrying know- >nothings whose sole intent with respect to >computers is to infect them with virii after >using them to send death threats to government >officials, find little kids to stalk, and look >at hardcore porn.
Ever sysadminned in a educational environment? You're pretty much described the exact problems that anyone looking after computers in a school/university has to deal with.
You forgot setting up hardcore porn http/ftp servers, IRC hijacking, and stealing the fscking mouse balls.
If you're not a Nazi about your computer labs, then Very Bad Things tend to happen. As you say, it's a few people ruining it for everyone.
It sucks for the admins, too, y'know. Most of them don't want to be assholes, but it's sheer self defence.
Slashdot Mirror
Buy 1 lab worth of AMD machines, and see how it works out. You've probably got a dozen different models of PCs out there anyway; one more flavor isn't going to hurt you too much.
If things go badly, much better to find out with 40 machines then 400.
>The way my company gets around this is when I
>have training, the company pays for it. But if I
>leave within 6 months of having the training I
>pay a portion of the cost: all the price if I
>leave immediately, 5/6 if I leave 1 month after
>the training, 4/6 after 2 months, etc. This way
>they get value for their money, and I get to
>leave if I want, but I have to pay my dues!
If someone else really wants to hire you away, they'll gladly pay that bill for your training.
That sort of thing is more of a speedbump then a barrier to a new job.
Why bother scanning email for porn when you could just scan the cache of your web proxy every few hours? In my experience there's a lot more porn coming in over port 80 then over the email system.
I have no idea about the content of your site, but I see a lot of sites where everything is dynamicly generated, for no obvious reason.
Maybe you could save a few thousand hits on your database by making some of the pages static, or just regenerating them with a perl script once an hour or something.
>Use their example, say, a spare part for your
>dishwasher. An 'L' pipe connector for instance.
>Your L connector break? Go to www.maytag.com,
>download the instructions for your 3D printer to
>construct a new part! Manufacturer's construction
>costs are eliminated!
Yeah, but Maytag will never go for it because spare parts are a revenue source for them. I'm sure they make a ton of money selling "L" pipe connectors at 200% markup.
>How does this program tell the difference between
/home because you don't really care.
/etc/fstab has been modified, it (hopefully!) won't be a surprise. It's the file changes you can't account for that you're supposed to worry about.
>an intruder modifying files using a real/spoofed
>login and a normal user modifying his own files
>that he should be modifying? Or is this program
>not designed to catch that?
Not familiar with this particular s/w, but with this sort of thing you can generally pick and choose which files/directories to watch. You're not going to bother checksumming
And if you're the admin, you're going to remember what you did. If you add a new HD or something and get an alert the next day saying that
My recommendation would be The UNIX programming Environment, Kernighan & Pike, ISBN 0-13-937681-X
It's what was used to teach the intro UNIX/C course that I took, it's an excellent book that I keep close at hand.
It's a bit dated though, you'll probably want to teach an editor other then ed(1).
>the DOS and Windows history doesn't show DR DOS,
>nor Desqview, nor VisiOn, nor Novell Netware,
>nor the IBM OS/2 history...
Novell Netware has nothing to do with DOS, and never has.
>The only way to do it is to have a group of
>people who are familiar with the ports systems of
>the respective bsd systems, BUT that doesnt hold
>any strong arogance towards a specific BSD. The
>project must be done independantly and not tied
>to a specific BSD.
I saw something here a few months ago about some Debian folks playing with the idea of a ports infrastructure for Debian... Might as well put them into the Grand Unified Ports Project too.
That's the whole thing, finding a bunch of people with a) the technical know-how, b) the time, and c) the diplomatic skills to pull this off.
Testing would be a cast-iron bitch: I can only imagine what could go wrong on all of those OSs and platforms.
I hate to sound cynical, but this is going to be one of those really good ideas that comes up every 6 months or so and dies down again because nobody is ambitious (crazy?) enough to tackle it.
(I'd love to be proved wrong, though.)
>If the problem still looks bad when I get back to
>it, then the next thing I do is ask a colleague
>for advice. In situations where there isn't
>another programmer around, I simply describe the
>problem to any willing listener. Often the simple
>act of talking about a problem elucidates the
>solution. Now this is going to sound silly, but
>I've even talked to my cat when working at home
>and sometimes had that "eureka" moment.
I heard of a university help desk who used a teddy bear as their first line of support; anyone with a programming problem had to explain it to the bear before a human would deal with them. Apparantly the bear solved lots of problems, too.
No, I don't remember where I heard this story, maybe it's just an urban legend.
>Know where the UPS keys and manuals and call
>sheets are. Laminate them and tape them to the
>wall, right by the flashlights (more that one)
>that you check once a week.
Hey, I thought I was paranoid!
(That's a compliment, of course!)
>Question: What are the options for a small-to-
>medium server room? (about 15-30 middling sized
>computers) Right now we have a single largish
>UPS, which says it's running at 40-60% of
>capacity. (depending on if I have the staging DB
>powered up) As we gain 2 more of those systems,
>and several other power-hungry systems, I know
>that our current UPS will not provide more than
>a minute or two of power for everything. We are
>not even monitoring it at the moment (we just
>hope that the power doesn't go out).
UPSen are rated in terms of KVA, and most of your equipment will probably have some info on the power supply stating it's power requirements.
Or you can bring in your friendly neighbourhood UPS salescritter, they have lots of practice making these calculations.
Monitors are pigs, if you can consolidate a few monitors with a switchbox then you'll save surprising amounts of power.
It's been a while, but here's what I remember...
MS Poxy supports the usual CERN http/ftp proxy stuff.
MS Poxy also has a proprietary proxy protocol that only works with their (vile) Windows client.
MS Poxy also supports Socks 4, but not Socks 5. Ask your admin guys about the settings, and try and find a Socks 4 client for whatever you're doing. I've gotten Mirc to work over Socks 4, so it can be done...
The proxy documentation is probably hidden somewhere on the MS web site, or if your admin guys are co-operative it's also on the CD.
>Is FreeBSD reasonably easy to install and try out
>in the same way that Linux is?
Yeah, the installation is comparable to the Slackware install, i.e. character based menus and keyboard navigation.
>Yes, I could write a bash script or perl script
/bin/sh and /bin/rm don't they?
/w Outlook's design, but I'll never say "that could never happen here" because it's too much like tempting fate.
>that deletes files. Guess what, not everyone uses
>bash and has perl on their unix system, and if
>they did, it would only delete their user files,
>and NO system files would be affected.
Well, you could find a *nix common denominator, i.e. pretty much everything has a
My 2 cents: we shouldn't get too complacent about this type of thing. Another nasty *nix worm/virus of some sort is probably inevitable. The more that people celebrate the problems that MS is having right now, the harder we're all going to have our noses rubbed in it when the next Morris worm comes around.
I'll gladly criticize problems
>What is the difference between a version (BSD)
>and a distro (Linux)?
The various Linux distros all have the same Linux kernel (maybe different versions, but it's all the same kernel, more or less).
Linux distros differ mostly in userland stuff, i.e. how the directories are laid out, init scripts, what utilities are bundled, packaging infrastructure, etc.
The various BSDs have different kernels. They all started from the same codebase, but have diverged significantly since.
Of course, the various BSD userland stuff has some differences as well.
>The Chapter Twenty Years of Berkley Unix: From
/catalog/opensources/book/kirkmck.html
>AT&T Owned to Freely Redistribuatble by Mcusick &
>Co. does a very good job at describing the
>movement and development from the AT&T based
>Berkley Unix to the Free versions that we have
>today.
The full text of this article is available at http://www.oreilly.com
>I think, though, that the reason Linux can have
>such a frequent release schedule is that it
>doesn't have a marketing behemoth behind it. It's
>target audience is largely comprised of
>technophiles who can pick up a new release, apply
>it correctly and troubleshoot the possible
>interoperability problems that arise with
>existing installed software. The Windows family
>is largely aimed at users who can't.
>The vehicle for distribution is different, too.
>The fact that Joe Windows User buys a box of
>software means that he does not want to have
>to buy another box of software in six months when
>the "new release" (read: service pack) comes out.
>Linux is mostly downloaded. Download the product.
>Download the updates. Because Windows is
>packaged, vast amounts of money would be wasted
>on repackaging for every service pack while older
>"releases" sat on the shelf never to be bought.
>(Example: do stores stock Windows95 next to
>Windows98 boxes? If they did, would anyone buy
>95?)
You're right of course, it's a big cost for the manufacturer and distribution channel when those boxes sit on the shelf and never sell. Red Hat etc. get burned by having surplus inventory as well, I still see some RH 5.2 boxes sitting around the local computer store.
One of the great things about FreeBSD is that there's a new release every 6 months. Most of the time it's just new drivers and teeny bug fixes, i.e. there was nothing too dramatic between 3.3 and 3.4.
But these new releases give me a bootable install CD that includes ALL of the latest h/w support, which greatly simplifies installation. No more piddling around adding SCSI drivers half way thru an NT install, or trying to build a custom install diskette, etc. etc.
Of course FreeBSD doesn't have any retail market to speak of, so a new rev doesn't cost them anything.
>How many patches, service packs, etc do you have
>to apply to WinNT4 server and IIS4 before they
>are secure enough to use?
Exactly, in a sane world we would have had NT 4.1, NT 4.2, etc instead of downloading 80MB+ service packs.
A lot of stuff changed with those service packs, it would have been worth spending $$$ on NT 4.1 just to get documentation that was up to date.
>These computers can be useful, providing nice
>email access and such, but are they really that
>good in the long run? Is this where computers are
>heading? If not, why are we arranging for our
>children to learn on these things, and not on
>real computers, which will undoubtedly have more
> challenging interfaces,
I don't think history is on your side here. The average PC user in 1992 knew how to move files around at a DOS prompt, was capable of formatting their own diskettes, etc etc.
This is certainly no longer the case. People are working around the clock to make interfaces as idiot-friendly as possible. (I don't think they are succeeding, but they sure are trying.)
>will be more capable, and will also be more
>complex. As a network administrator, I see
>users who cannot understand the simple (in my
>mind) concept of a directory tree structure.
>Shouldn't we be focusing on teaching our kids
>how to use actual computers? Unless, this is
>where computers are heading. I like to think not,
>however the current trend appears to be
>simplification of computers, rather than
>education of users. I think those two are both
>needed, but they should meet in the middle
>somewhere. Whatever. Random mumblings from a
>lunatic, take them as you will.
I agree with you. If you can't understand what a directory structure is, then go use a fscking typewriter or something. The industry, and the average consumer, have different ideas, I guess.
This little puppy would have an advantage for schools, in that it's impossible to screw up the software configuration. A few labs full of tamper-proof machines would be worth hundreds of man hours a year.
>Anyone thinking this is happening very close to
>april 1? I just cannot fathom the ultima series
>without its original designer.
They (Origin) have been saying that Ultima IX was the last Ultima.
I'm shedding no tears, Ultima has been going steadily downhill. Those guys clearly had no intention of making another U5 or U7; I don't know if that's Garriot's fault or the marketroids'.
Hopefully a fresh start will give Garriot a chance to recapture the magic of those first few games. I admire the guy, he's clearly brilliant, but those last 2 "Ultima" games don't deserve the name.
>>Kittens direct to the desktop.
>That's what the internet is all about: Terrified
>kittens hurled at supersonic speeds.
>It brings a new slant to the term fat pipe. How
>exactly do you measure that bandwidth ? kps
>(kittens per second?)
Exactly. Make sure that all your cable is up to CAT5 standards, or you might experience purrity errors.
>Hmm. Much though I dislike this tactic in
>Microsoft, it is probably worth pointing out that
>Novell, M$'s biggest competitor in the network
>server market, are doing almost exactly the same
>thing to their CNEs - "upgrade to a Netware 5
>CNE, or lose it"
Sort of, not really.
NW5 has been out since Fall 98, and you have until Aug 2000 to recertify. This requires 1 test for 4.x CNEs.
W2K has been out since Feb 2000, and you have until Dec 2000 to recertify. This requires 2 or 3 tests IIRC.
Which do you think is a bigger rip off?
>Someone stole ALL the mouseballs from one of the
>computer labs.
Someone did that at my Uni a few years back.
So the Computer Services guys glued ALL of the mice shut. In ALL of the labs in that building. For all I know, the Uni is still glueing it's new mice shut.
So there are hundreds of mice that can't be cleaned and are slowly getting gummed up and unusable... All because of some asshole prank.
>Furthurmore, someone has everyone believing that
>all students are computer virus carrying know-
>nothings whose sole intent with respect to
>computers is to infect them with virii after
>using them to send death threats to government
>officials, find little kids to stalk, and look
>at hardcore porn.
Ever sysadminned in a educational environment? You're pretty much described the exact problems that anyone looking after computers in a school/university has to deal with.
You forgot setting up hardcore porn http/ftp servers, IRC hijacking, and stealing the fscking mouse balls.
If you're not a Nazi about your computer labs, then Very Bad Things tend to happen. As you say, it's a few people ruining it for everyone.
It sucks for the admins, too, y'know. Most of them don't want to be assholes, but it's sheer self defence.