What you describe is the common configuration, where one developer becomes a code primadonna. He (or she) will be the only one who really understands the system and everyone else just try to work around the primadonna and avoid getting in his or hers way.
Then again, this guy has done 38 man-years of work in one year.
Such behavior is just a level of refinement away from the brilliant social satire done by Michael Moore [imdb.com], the genius behind "TV Nation" and "The Awful Truth".
On the other hand, it's just a level away from the work of Tom Green, too, whom I won't dignify with a link.
Anybody can walk around with a camera and act like an asshole. Saying "I'm doing it for artistic reasons" doesn't make it art, unless you also think Yoko Ono scrawling "fuck" on a museum ceiling is art.
This guy has a valid point, but the only people who are going to listen long enough to hear it are those who already get it.
Probably interesting, but entirely too long to read. Its the holidays - I'm not in the mood for pages and pages of stuff about law, I'm in the mood for candy canes...
Bookmark it and read it Wednesday. Merry Christmas.
The original point of this article was Free Software, and how administrative cost is the most significant aspect of system operations. Since software updates and user additions are a part of _ANY_ coputer system, and are not specific to a Microsoft or Free Software solution, why are you involving them?
I'm not; I'm responding to a post by someone who claimed that if you have machines with long uptimes, no maintenance will be required.
I was showing how a school doesn't just have servers, it has desktops that must be maintained as well. I was further showing that long uptimes are deceptive, because even if every individual machine runs for a very long period of time, a large group of machines will have frequent failures, not all just go out at once.
MTBF is just that; MEAN time. As in average. Not "guaranteed exact time of failure."
If you have 2 years of MTBF, as the poster implied, and you have 24 systems, that's a failure a month. Long individual uptimes don't mean no maintainance is required, they just tell you how much workload will be likely for maintainance.
Many schools make the mistake that poster is making; assuming that it is possible to set systems up so that they'll require no maintainance. It isn't; not with the money schools can spend.
Everybody's accusing me of being anti-Free-Software here, but nothing could be further from the truth; I believe that the best thing schools can do is outsource their computer maintainance, and instead of saying "we want Windows" they should say "we need to do this and this and this" and let the outsourcer set up the best way to acheive those goals in a maintainable manner. I'm confident that will often involve free software, and when it doesn't it will often involve Open Systems (I.E., things like a Sun server and a bunch of SunRay thin clients), and only rarely a Windows-based solution.
I make my living running non-Windows solutions, it's laughable in the extreme to call me pro-Windows. Hell, I'm making my wife buy a laptop so I can take Windows off the last remaining workstation in my house that dual-boots it.
In your second link, the very first part of it says I'm right; if you have two disks eith an individual MTBF of 10,000 hours, you'll have a failure every 5,000 hours on average. Which is exactly what I said, only I was using 200 units with an individual MTBF of 56,000 hours.
I pulled those numbers out of my ass as examples, but the math was 100% correct, according to the very links you quoted as rebuttal.
Uhhh...so how is this different from Windoze machines? Are we trying to compare the two or not?
Different in the number of people it takes to maintain it correctly. That's about it.
Oh, it's easier to maintain the Linux machines from off-site, but it's doable with Windows too.
The key is the fact that it takes fewer people to maintain the Linux systems, and that it's easier to set them up so that the kids can't screw them up.
But with either, it's going to take outsourcing for the smaller schools to afford to do it right. Right now they're mostly doing it wrong. Mostly.
When you have servers that have uptimes of two years (or more), something tells me that some of those servers are, more or less, a fire-and-forget system. Anybody who sets up a server right WON'T need to maintain it much.
As long as there are no users being added, no programs being added, and you only have a handful of systems, you might be right.
But let's say you have 200 systems, with a mean time between failures of 56,000 hours each.
That's one failure every 12 days, more or less.
A school has dozens or hundreds of systems, with much shorter MTBF on the physical hardware, and has hundreds of students using those machines. They require security monitoring, hardware replacement, software configuration and upgrading; near-constant attention, if it's larger than one server and a handful of clients.
If I take any one of my servers and point at it and base my manpower computations on that server alone, the numbers will look deceptively like I can do it all myself. When I broaden my sights out to all of the several hundred large servers I manage, I instead get a 7-man team rotating on-call duties between 3 production and 5 test projects, and the thought of doing it all myself becomes laughable.
A typical school is somewhere in the middle if you want to use computers for education, instead of (as I said) sticking a few PCs in the physics lab and letting the brightest students do WTF the want with them.
If you just want to stick a file server in the secretary's office and put a PC on each of the administrator's desks, you're probably right. But I'm talking about a school using computers for educating the kids, not a school using computers near the kids.
A good computer system does not need a lot of maintenance.
Horseshit. Any non-trivial setup needs quite a bit of maintainence, and a school needs a non-trivial setup if you want to get any non-trivial use out of it. Otherwise it's doorstop PCs in the wings of the physics class getting used by half a dozen kids a year.
The answer for small schools is going to have to be outsourcing.
Yeah, terrorists are slamming crowded jetliners into crowded skyscrapers, some nut is sending deadly diseases in the mail, the economy is in a recession that's putting thousands of our fellow geeks out of work, and Americans are dying in a war, but at least we've got neat video games.
Upon re-reading what I wrote above, I realize I said something misleading.
With Ximian you don't pay for updates. You only pay for the fastest-available access to them. Updates are still free.
With RedHat Network, you pay if you want to avoid having to manage your entitlements via a web page if you have multiple systems. You get the same bandwidth priority either way.
I have no problem with them trying to find ways to make money, but this one isn't going to work.
Let's say you're using RedHat. It'll be a similar story with other distributions, but that's the one with which I'm most familiar.
With Ximian, you get GNOME slightly ahead of what RedHat has, major hassles with upgrading RedHat to a new version, and you pay for updates.
With Redhat and no Ximian, you get GNOME slightly behind the curve, easy upgrades to new versions, and updates are free but not quite as easy to use as Ximian.
So basically, I'm supposed to pay a monthly fee so that I can have GNOME be the most recent build, instead of a couple months old? Please; if I'm in a position where that's really important, I'm someone who can fix it myself.
I don't think that many people are going to find value in this particular service.
Oh; and I should add, I hope I'm wrong. I wish no ill-will to Ximian, and I'd love to be wrong about this.
Slashdot Mirror
Let me quote: "In March, just 14 companies controlled 60 percent of users' online time, down from 110 companies two years earlier,"
So there is actually *LESS* competition than before as fewer and fewer companies control the web.
And yet, 1.5Mb access costs $40-$50 a month now, whereas the best you could get five years ago was a couple grand a month for the same speed.
Damn those economies of scale, they're ruining the internet!
The commercialization of the internet has given rise to free web page services
There; I edited your comment to bring out the salient point.
Yes, the government often speaks out of the side of their mouths. However, this is blatant and obvious miscoduct.
:-)
There should be some inquiry into this matter.
Moderators: If you have to look up any of the terms I've used, don't moderate me. You're probably confused.
So, if we don't know what "miscoduct" is, we can't moderate you?
but he did do some societal good.
:-)
I don't recall saying he did no GOOD; I said he did no ART.
A sewer does societal good, too, but I ain't hanging what comes out the other end up on my wall.
Hell, some guy owes him a debt of gratitude for making Drew Barrymore all weepy and vulnerable, too.
What you describe is the common configuration, where one developer becomes a code primadonna. He (or she) will be the only one who really understands the system and everyone else just try to work around the primadonna and avoid getting in his or hers way.
Then again, this guy has done 38 man-years of work in one year.
"You are the product of a mutational union of ~640Mbytes of genetic information."
I hate to respond to siglines, but does that mean that 640M is enough for any person?
Such behavior is just a level of refinement away from the brilliant social satire done by Michael Moore [imdb.com], the genius behind "TV Nation" and "The Awful Truth".
On the other hand, it's just a level away from the work of Tom Green, too, whom I won't dignify with a link.
Anybody can walk around with a camera and act like an asshole. Saying "I'm doing it for artistic reasons" doesn't make it art, unless you also think Yoko Ono scrawling "fuck" on a museum ceiling is art.
This guy has a valid point, but the only people who are going to listen long enough to hear it are those who already get it.
I'll use QuickTime on both my PC AND my Macs
"We got both kinds of music; Country AND Western."
QuickTime is pretty damn standard on every platform but Linux.
Two platforms is not "every platform except Linux".
Show me the Solaris viewer and I'll use that.
It is only in .mov format so far, so mpg linux people will have to move to a windows box until they post the other formats.
No, we won't "have" to do any such thing.
If they don't care enough to use a standard format, I certainly don't care enough to view their movie.
I think it's even worse than how you present it:
A witness says that Al-Queda deliberately set out to leave back doors and security holes in XP.
XP then has the worst hole of any Microsoft OS, ever.
The FBI suddenly has a lot of questions. They damn well should.
Probably interesting, but entirely too long to read. Its the holidays - I'm not in the mood for pages and pages of stuff about law, I'm in the mood for candy canes...
Bookmark it and read it Wednesday. Merry Christmas.
Gnomehack.
The original point of this article was Free Software, and how administrative cost is the most significant aspect of system operations. Since software updates and user additions are a part of _ANY_ coputer system, and are not specific to a Microsoft or Free Software solution, why are you involving them?
I'm not; I'm responding to a post by someone who claimed that if you have machines with long uptimes, no maintenance will be required.
I was showing how a school doesn't just have servers, it has desktops that must be maintained as well. I was further showing that long uptimes are deceptive, because even if every individual machine runs for a very long period of time, a large group of machines will have frequent failures, not all just go out at once.
MTBF is just that; MEAN time. As in average. Not "guaranteed exact time of failure."
If you have 2 years of MTBF, as the poster implied, and you have 24 systems, that's a failure a month. Long individual uptimes don't mean no maintainance is required, they just tell you how much workload will be likely for maintainance.
Many schools make the mistake that poster is making; assuming that it is possible to set systems up so that they'll require no maintainance. It isn't; not with the money schools can spend.
Everybody's accusing me of being anti-Free-Software here, but nothing could be further from the truth; I believe that the best thing schools can do is outsource their computer maintainance, and instead of saying "we want Windows" they should say "we need to do this and this and this" and let the outsourcer set up the best way to acheive those goals in a maintainable manner. I'm confident that will often involve free software, and when it doesn't it will often involve Open Systems (I.E., things like a Sun server and a bunch of SunRay thin clients), and only rarely a Windows-based solution.
I make my living running non-Windows solutions, it's laughable in the extreme to call me pro-Windows. Hell, I'm making my wife buy a laptop so I can take Windows off the last remaining workstation in my house that dual-boots it.
In your second link, the very first part of it says I'm right; if you have two disks eith an individual MTBF of 10,000 hours, you'll have a failure every 5,000 hours on average. Which is exactly what I said, only I was using 200 units with an individual MTBF of 56,000 hours.
I pulled those numbers out of my ass as examples, but the math was 100% correct, according to the very links you quoted as rebuttal.
Uhhh...so how is this different from Windoze machines? Are we trying to compare the two or not?
Different in the number of people it takes to maintain it correctly. That's about it.
Oh, it's easier to maintain the Linux machines from off-site, but it's doable with Windows too.
The key is the fact that it takes fewer people to maintain the Linux systems, and that it's easier to set them up so that the kids can't screw them up.
But with either, it's going to take outsourcing for the smaller schools to afford to do it right. Right now they're mostly doing it wrong. Mostly.
Ok, I'll bite. How do YOU calculate it?
When you have servers that have uptimes of two years (or more), something tells me that some of those servers are, more or less, a fire-and-forget system. Anybody who sets up a server right WON'T need to maintain it much.
As long as there are no users being added, no programs being added, and you only have a handful of systems, you might be right.
But let's say you have 200 systems, with a mean time between failures of 56,000 hours each.
That's one failure every 12 days, more or less.
A school has dozens or hundreds of systems, with much shorter MTBF on the physical hardware, and has hundreds of students using those machines. They require security monitoring, hardware replacement, software configuration and upgrading; near-constant attention, if it's larger than one server and a handful of clients.
If I take any one of my servers and point at it and base my manpower computations on that server alone, the numbers will look deceptively like I can do it all myself. When I broaden my sights out to all of the several hundred large servers I manage, I instead get a 7-man team rotating on-call duties between 3 production and 5 test projects, and the thought of doing it all myself becomes laughable.
A typical school is somewhere in the middle if you want to use computers for education, instead of (as I said) sticking a few PCs in the physics lab and letting the brightest students do WTF the want with them.
If you just want to stick a file server in the secretary's office and put a PC on each of the administrator's desks, you're probably right. But I'm talking about a school using computers for educating the kids, not a school using computers near the kids.
A good computer system does not need a lot of maintenance.
Horseshit. Any non-trivial setup needs quite a bit of maintainence, and a school needs a non-trivial setup if you want to get any non-trivial use out of it. Otherwise it's doorstop PCs in the wings of the physics class getting used by half a dozen kids a year.
The answer for small schools is going to have to be outsourcing.
This is supposed to be the anti-Trek, right?
:-)
So, what, you'll play a character the fans like?
Come on... not to troll or flame, but these shows only have popularity cause of one thing:
Great looking chicks.
Then why aren't other shows with MORE hot chicks more popular with the same audience?
Buffy draws an audience that includes a lot of straight girls. Dawson's Creek isn't very popular with straight guys 24-45.
In short, "not to troll or flame" was bullshit, you're trolling.
Actually - with RHN you get the updates WAY before Ximian are putting them in their Red Carpet servers.
True. That is in fact one of the two primary reasons I stopped using Ximian.
The other was a bout of upgrade hell when I tried to upgrade a RedHat 7.0/Ximian system to RedHat 7.1 without first removing Ximian entirely.
Yeah, terrorists are slamming crowded jetliners into crowded skyscrapers, some nut is sending deadly diseases in the mail, the economy is in a recession that's putting thousands of our fellow geeks out of work, and Americans are dying in a war, but at least we've got neat video games.
Upon re-reading what I wrote above, I realize I said something misleading.
With Ximian you don't pay for updates. You only pay for the fastest-available access to them. Updates are still free.
With RedHat Network, you pay if you want to avoid having to manage your entitlements via a web page if you have multiple systems. You get the same bandwidth priority either way.
I have no problem with them trying to find ways to make money, but this one isn't going to work.
Let's say you're using RedHat. It'll be a similar story with other distributions, but that's the one with which I'm most familiar.
With Ximian, you get GNOME slightly ahead of what RedHat has, major hassles with upgrading RedHat to a new version, and you pay for updates.
With Redhat and no Ximian, you get GNOME slightly behind the curve, easy upgrades to new versions, and updates are free but not quite as easy to use as Ximian.
So basically, I'm supposed to pay a monthly fee so that I can have GNOME be the most recent build, instead of a couple months old? Please; if I'm in a position where that's really important, I'm someone who can fix it myself.
I don't think that many people are going to find value in this particular service.
Oh; and I should add, I hope I'm wrong. I wish no ill-will to Ximian, and I'd love to be wrong about this.