What's the translation between hits and pages served? It's not 1:1, or anything close to that. Each graphic button is a separate hit, etc.
I give/. an amazing hand from a technical, a journalistic, and a community viewpoint.
=Blue(23)
Re:$ is made from support contracts!
on
IBM Wants Linux
·
· Score: 1
Well, since IBM already sells Linux support contracts (stopped by their booth at last years AIX Technical Univeristy), they are gearing up for this, too.
Still in it's infancy though - they didn't have anything in place for global support, etc. Have a way to go until it's the same as their AIX support.
From what I understand, Deep Blue was designed to play specifically against Kasparov's play style. Deep Fritz doesn't have that advantage. Normally chess masters can beforehand analyse each other's styles - Kasparov couldn't do that with Deep Blue. That can be done with Deep Fritz who has many games on record. As much as I'd love to bet on the computer, my money would be on the chess master.
Notice how IBM dismantled Deep Blue. My thoughts on the matter are that it did what they wanted it to do, got the publuicity, and they didn't want a chance for it to be beat.
To be honest, I wouldn't ask too many questions on specific commands - that's what man pages are for, and they might have been focusing on other things.
One of the big things I would check for is troubleshooting skills. And in a non-obvious way, so they don't zero in on what you're asking for and give the "right" answers. Asking to give an example of a problem in the past and what they did, give some hypothetical situations (though some people think better in front of a keyboard then in when speaking.)
On big one with me is automation and tools. I don't care if you know a specific tool - that can always be learned. But once you get to real sizes, you need to use automation and tools, you can't do everything by hand. If you told me that you speced out or even wrote tools to fit the specific circumstances of the last job, that's a big plus. Along the same lines, any sysadmin that can't take the time to be fluent in a shell probably isn't worth my time. I ask them for their prefered shell and why. It doesn't really matter what they answer, as long as they have an answer. Along those lines, tellign me that "they used to love [insert shell], but now they don't care as much because they always use perl (or other appropriate language)" is also fine.
Sysadmining is sometimes periods of boredom followed by periods of extreme need. If you can keep your cool in that extrene need, that's very good, but hard to judge on an interview. It's very important, though. If you're a self-starter, and those periods of boredom will be used on projects to make your job easier, either from a manager or self-starting, is also good, and something that might be easier to detect in an interview.
Many sysadmins have a large (and fairly well-deserved) ego. This is almost a "necessary evil". However, a prima-donna or someone who will not work with other team members is a problem, and that can be determined to a point during an interview. Also watch out for loose cannons. They can be great, but they're hard to control. A small company might benefit more then a large one by a loose cannon, but no matter how good they are they can get you in trouble. You just need to balance if it's worth it.
I'm not sure that a Window's box is quantum. For example, whenever I observe one, the superposition of UPTIME vs. DOWNTIME collapses in the same direction. Always.
I'm a Unix Sysadmin, and salaried. One company I worked for was no pay for on-call. Sucked, but it wasn't a big company. Company I'm working at now used to have no policy, but the managers were "forgiving" - if you were up late working on an on-call problem, they didn't hassle you if you came in late. And you weren't front line - the front line help-desk hourly people did get a bonus any week they carried the duty pager.
Since we're in multiple time zones, but have only one primary IT headquarters, things are crazy. As things have gotten more hectic (and some people left because of the unpaid on-call time) they've revamped the schedule. Now we have more covered hours by in-office 1st line phone firewall, there is a bonus for being on-call (though still no money for time spent doing things), and finally if you do get called when you're not on call, IF you answer you get a flat per incident fee. Quite generous, and it's gone a long way towards building the morale of the same people that used to be burnt out.
I especially like the part where they encrypt the data, so that it can't be monitored without breaking the DMCA. However, the "wrong to open a file you downloaded" doesn't sound like it will hold up in court. It is the nature of files to be manipulated.
And they are right - that has a good deal of non-infringing uses.
Slightly off topic, there is an urban legend that grapefruit juice breaks down the enzime that gives you "caffeine tolerance", so you can get around it for a little bit until your body makes more.
IANABC. I am not a biochemist. 8)
=Blue(23)
What do WE think the system of 2010 would be?
on
The Computer of 2010
·
· Score: 1
Ok fellow/.ers, we're all picking this apart like crows on a corpse. Let's look on the flip side - what the heck do we think the "computer of the 2010's" will look like?
Here a couple of suggestions to get us started, feel free to disagree - this is just one concept:
1. We won't know what it looks like, 'cause we'll never see it. We may have a couple of interfaces to it, some looking like PDAs, monitors, keyboards, clipboards, whiteboards, etc. But none of them are the actual machine, they all just connect to it wirelessly. Either with lots of bandwidth, compressed data streams, or both. And don't forget encrypted.
2. Hard disk? What's that? Maybe we'll still have it for our larger (>1 TB) needs, since that much memory might be expensive in terms on money or power consupmtion. And since we're not carting around the actual system, it doesn't matter as much if it's a moving part or not.
3. CPU. Huh? Why one, and why even a central one? Personally, this whole rush for a faster chip seems kinda silly considering SMP & such. However, those don't scale as well unless they have work for all the processors. Do we expect lots of good stuff for theat by 2010? Sure, why not, this is my fantasy. Anyway, if they are runnign my house, my office, my pr0n surfing, and my investments, my ray tracing,..., I think we can keep 'em busy. So maybe 16-way 10GHz system for the newbies, with us serious users going more.
Which of course, brings us to speciallized PUs. Right now we have our FPUs, our graphics PU, etc. Let's have lots of them, which we can add on to have the system that is right for us. The voice-recognition folks have their VPUs, the 3D gamers have their FPUs & GPUs (oh wait, they already do), etc. Whatever we need.
3. Interface - whatever we want. As long as we can define data streams, we have interfaces. Some are output only (printers, they'll still be around), some may be input only (keyboards), some may be both (touch sensitive screens, speach). Why tie the computer to the interface at all? As long as their are "drivers" to render the input into a form the computer likes it, it should take input from your VR glove as easily as your smartpen as easily as your chord keyboard.
What other ideas do we have? Is this worthy of a "Ask Slashdot?" 8)
Muhahaha, since Episode I stunk so bad, those of us who believe that screwing up once doesn't mean you've lost it completely will be able to get tickets without standing on line.
Which of course brings us back to the real title for Episode II: Revenge of the Jar Jar.
This title will scare away so many of you "non-believers" that scoring tickets will be easy. Of course, it's all a fake - those of us in the know understand that it has nothing to do with Jar Jar, we'll be laughing up our sleeves the entire time.
Is that rebate anything like the Compuserve 2000 "rebate" in the states, where they give you $400 off on a machine, if you spend much more then that on their service? 8)
Only $45 after $2,000,000 mail-in-rebate*
*Rebate requires 41,400 UPC codes from specially marked "Linux for S/390" boxes of your favorite operating system. Stays crunchy in milk.
Hidden virii that mutate? You know, this would explain much about Windoze.
Think about it. An organism is much more likely to grow in an environment that provides it's needs. This is why your bread gets moldy but your car usually doesn't. (And if it does, why I'm sorry for you). But getting back to the topic, say that Win32 has lots of nice little niches for virii. Not really a stretch of anyone's imagination. Now assume that a self mutating virus would spread there better then someplace without all of the problems.
So now we have our Win32 petri dish. How do we test for it. Well, we can look for the virii themselves, but that may be hard, they are small. Let's look for some second order effect. Well, self-mutating virii would occasionally end up bad and crash either an app or the system. Hmm, check, windows has that. Some virii would mutate away from being able to save themsevles to disk, so time went on your system would get slower and slower from more virii, and rebooting would make things nicer for a time since all of those non-disk virii would go away. Windows - check.
That's good enough for me. Windoze is really a petri dish for a new flavor of virtual life. Next the EPA will be getting an injuction against people rebooting NT servers.
IBM is really working to keep the top. They've got another one on the drawing board that's a different architecture then this called SMASH (love that name!). It's for "Simple, Many, And Self-Healing".
Here's a link to an article, but's it's a bit dated: http://www.ibm.com/news/1999/12/06.phtml
I'm well versed with IBM's LVM, but have had no exposure to the Linux one. How to they compare?
I could see how IBM knows how their LVM works, and wouldn't want to "re-invent the wheel" trying to fit that into the existing Linux LVM mold. But that's only if at an underlying level they are incompatable. Can anyone compare them deep?
IBM's LVM does support true mirroring, and did 3 years ago. Was classified as real RAID 1. Perhaps you were thinking of their problem with mirroring and striping. They used to have a problem with that, and their striping didn't match RAID 0 requirements three years back.
Slashdot Mirror
What's the translation between hits and pages served? It's not 1:1, or anything close to that. Each graphic button is a separate hit, etc.
/. an amazing hand from a technical, a journalistic, and a community viewpoint.
I give
=Blue(23)
Well, since IBM already sells Linux support contracts (stopped by their booth at last years AIX Technical Univeristy), they are gearing up for this, too.
Still in it's infancy though - they didn't have anything in place for global support, etc. Have a way to go until it's the same as their AIX support.
=Blue(23)
Notice how IBM dismantled Deep Blue. My thoughts on the matter are that it did what they wanted it to do, got the publuicity, and they didn't want a chance for it to be beat.
=Blue(23)
One of the big things I would check for is troubleshooting skills. And in a non-obvious way, so they don't zero in on what you're asking for and give the "right" answers. Asking to give an example of a problem in the past and what they did, give some hypothetical situations (though some people think better in front of a keyboard then in when speaking.)
On big one with me is automation and tools. I don't care if you know a specific tool - that can always be learned. But once you get to real sizes, you need to use automation and tools, you can't do everything by hand. If you told me that you speced out or even wrote tools to fit the specific circumstances of the last job, that's a big plus. Along the same lines, any sysadmin that can't take the time to be fluent in a shell probably isn't worth my time. I ask them for their prefered shell and why. It doesn't really matter what they answer, as long as they have an answer. Along those lines, tellign me that "they used to love [insert shell], but now they don't care as much because they always use perl (or other appropriate language)" is also fine.
Sysadmining is sometimes periods of boredom followed by periods of extreme need. If you can keep your cool in that extrene need, that's very good, but hard to judge on an interview. It's very important, though. If you're a self-starter, and those periods of boredom will be used on projects to make your job easier, either from a manager or self-starting, is also good, and something that might be easier to detect in an interview.
Many sysadmins have a large (and fairly well-deserved) ego. This is almost a "necessary evil". However, a prima-donna or someone who will not work with other team members is a problem, and that can be determined to a point during an interview. Also watch out for loose cannons. They can be great, but they're hard to control. A small company might benefit more then a large one by a loose cannon, but no matter how good they are they can get you in trouble. You just need to balance if it's worth it.
=Blue(23)
=Blue
Since we're in multiple time zones, but have only one primary IT headquarters, things are crazy. As things have gotten more hectic (and some people left because of the unpaid on-call time) they've revamped the schedule. Now we have more covered hours by in-office 1st line phone firewall, there is a bonus for being on-call (though still no money for time spent doing things), and finally if you do get called when you're not on call, IF you answer you get a flat per incident fee. Quite generous, and it's gone a long way towards building the morale of the same people that used to be burnt out.
=Blue(23)
And they are right - that has a good deal of non-infringing uses.
=Blue
=Blue(23)
IANABC. I am not a biochemist. 8)
=Blue(23)
Here a couple of suggestions to get us started, feel free to disagree - this is just one concept:
1. We won't know what it looks like, 'cause we'll never see it. We may have a couple of interfaces to it, some looking like PDAs, monitors, keyboards, clipboards, whiteboards, etc. But none of them are the actual machine, they all just connect to it wirelessly. Either with lots of bandwidth, compressed data streams, or both. And don't forget encrypted.
2. Hard disk? What's that? Maybe we'll still have it for our larger (>1 TB) needs, since that much memory might be expensive in terms on money or power consupmtion. And since we're not carting around the actual system, it doesn't matter as much if it's a moving part or not.
3. CPU. Huh? Why one, and why even a central one? Personally, this whole rush for a faster chip seems kinda silly considering SMP & such. However, those don't scale as well unless they have work for all the processors. Do we expect lots of good stuff for theat by 2010? Sure, why not, this is my fantasy. Anyway, if they are runnign my house, my office, my pr0n surfing, and my investments, my ray tracing, ..., I think we can keep 'em busy. So maybe 16-way 10GHz system for the newbies, with us serious users going more.
Which of course, brings us to speciallized PUs. Right now we have our FPUs, our graphics PU, etc. Let's have lots of them, which we can add on to have the system that is right for us. The voice-recognition folks have their VPUs, the 3D gamers have their FPUs & GPUs (oh wait, they already do), etc. Whatever we need.
3. Interface - whatever we want. As long as we can define data streams, we have interfaces. Some are output only (printers, they'll still be around), some may be input only (keyboards), some may be both (touch sensitive screens, speach). Why tie the computer to the interface at all? As long as their are "drivers" to render the input into a form the computer likes it, it should take input from your VR glove as easily as your smartpen as easily as your chord keyboard.
What other ideas do we have? Is this worthy of a "Ask Slashdot?" 8)
=Blue(23)
By definition, all software sucks. Some just sucks more...
Which of course brings us back to the real title for Episode II: Revenge of the Jar Jar.
This title will scare away so many of you "non-believers" that scoring tickets will be easy. Of course, it's all a fake - those of us in the know understand that it has nothing to do with Jar Jar, we'll be laughing up our sleeves the entire time.
=Blue
Only $45 after $2,000,000 mail-in-rebate*
*Rebate requires 41,400 UPC codes from specially marked "Linux for S/390" boxes of your favorite operating system. Stays crunchy in milk.
Think about it. An organism is much more likely to grow in an environment that provides it's needs. This is why your bread gets moldy but your car usually doesn't. (And if it does, why I'm sorry for you). But getting back to the topic, say that Win32 has lots of nice little niches for virii. Not really a stretch of anyone's imagination. Now assume that a self mutating virus would spread there better then someplace without all of the problems.
So now we have our Win32 petri dish. How do we test for it. Well, we can look for the virii themselves, but that may be hard, they are small. Let's look for some second order effect. Well, self-mutating virii would occasionally end up bad and crash either an app or the system. Hmm, check, windows has that. Some virii would mutate away from being able to save themsevles to disk, so time went on your system would get slower and slower from more virii, and rebooting would make things nicer for a time since all of those non-disk virii would go away. Windows - check.
That's good enough for me. Windoze is really a petri dish for a new flavor of virtual life. Next the EPA will be getting an injuction against people rebooting NT servers.
Save the virii!
=Blue(23)
Here's a link to an article, but's it's a bit dated:
http://www.ibm.com/news/1999/12/06.phtml
=Blue(23)
I could see how IBM knows how their LVM works, and wouldn't want to "re-invent the wheel" trying to fit that into the existing Linux LVM mold. But that's only if at an underlying level they are incompatable. Can anyone compare them deep?
=Blue(23)
=Blue(23)