I was working for the IT department of an austrian company in Tokyo. The company grew and office space is expensive, especially if you need to move to another building. The simple fix was, to let some of the developers and translators, who were working already several years at the company, work from home. After all, those do not need physical appearance and they prefer (due to the nature of work) a quiet environment. Something which is difficult to get in a japanese company.
Worked out well, as it was easy to check they are working by checking the results. The employees (not all wanted to work from home) were generally happy, some office space was saved, travel money (paid usually by the company) was saved, in the end, everyone was happy.
I think the trick in this excercise was, so let experienced workers work from home. People who are known to be able to motivate themself. And as everyone could check the productivity, the usual problem of teleworking, not being able to tell if the employee watches TV or works 8 hours, did not apply here.
Re:Instead...Japan, the land of confusion
on
Making Change
·
· Score: 3, Insightful
Life gets more complicated when coming originally from Germany, where VAT is always included in the advertized price (for end-users, not for businesses) and going to Japan, where 5% tax is mostly added. About 90% of all times I need to add 5% at the cashier. In the other cases I don't need to. Now that makes calculation the price to pay complicated.
My experience reflects this too. While theoretically and actually practically (when doing large file transfers) the speed is enough, you get dropped frames as the data does not flow as a steady flow. Even a modest 1.5MBit/s divx is far from perfect when watching it being transfered from a (fast) file server to a (fast) computer. It's ok for short clips, but watching a movie is out of question. I prefer to copy the movie to disk and then watch it. Yet, I can move about 600kByte/s data on the 11MBit/s wireless LAN.
Wouldn't that be even better? This way they know what to expect from you and (hopefully) that will result in less spam from them. Simple calculation: this potential customer costs us more maney than we can possible earn from this him/her.
And since you do nothing illegal, no need for fear! In the easiest case send them their empty envelope back. You can always say that you forgot to put something in.
It's a nice way to check a connection is not made, that packets do not go out of one or another interface, that traffic is encrypted. tcpdump can do the same (except follow TCP traffic, which is very enlightning for users who like telnet).
So while Ethereal does not increase security by itself, it does add security by making it possible to check out the packets. That makes is IMHO a security tool.
Here in Tokyo, the place where you want to be when you want Internet connections for cheap, the standard DSL service is 12Mbit/s down, 1Mbit/s up. For abour Yen 3000 (about US$ 26). And so far no restrictions. And it's fast (within Japan 900kbyte/s if the server is fast enough, to USA usually 200kbyte/s).
Everything else in Japan and especially in Tokyo is expensive. But Internet is as cheap as you can imagine.
Starting at Ultra320, Quick Arbitration is now available.
Given that this is cutting edge of parallel SCSI, I can understand Sun to skip anything older. However I yet have to see a significant performance gain from going from U2W (80 MB/s) to FC. Arbitriation might be slower on U2W, but FC contains routing informations in each packet, which parallel SCSI lacks.
Personally, I am very happy with the good old parallel SCSI. Even cutting edge drives like Cheetah 15k.3 are really fast, even when I cannot push them to their limit.
Houses that are supposed stay cool in the summer should be well insulated.
That's what I thought and that's what common sense is. But not here in Japan. Summer is humid and long (not just 1 week humid hot like it is usually in Germany once in a year). 3 Monthes humidty of easily 70% up to 90% and temperatures of 30+ centigrades. At night it often stays between 25 and 30 centigrades.
Now imagine a cool inside...you can nearly feel the mold growing. It would be ok if you never need to leave the house, but open the door and, baaam, humidity comes in.
So the perfect solutions would be something that keeps the inside insulated from the outside (temperature wise), yet allows a free flow of air. That seems to be tricky. I have not found a solution.
Old houses in Japan have a clever air circulation system: the top of the room walls are open (can be closed in winter) and as hot air moves up, it's pushed out by a slight breeze. Many of the outside walls can be opened using sliding doors. So if there is any breeze outside, it's not too hot inside. 30 degrees outside or not.
Unfortunately it became unpopular to have the top part of room walls open, as there is no privacy any more. Any noise in a house is audible in all connected rooms. And air conditioners completely changed the way of cooling, introducing the humidity problem. And all this is amplified by having a lot of concrete, increasing the outside temperatures even more on a sunny day. Most days in summer are sunny. That's modern Tokyo. But that's another problem.
Modern houses in Japan are really built to last 20 years. Maybe 30. I live in one which is 16 years old and the owner probably would like to rebuilt it (which is not soo expensive as one might think, as the ground is a very expensive part of a house, and the house itself is built by "wood and paper and plastic". Not many stones here...
Any idea why there are 0 of them in Japan? Japan runs on 50Hz east side and 60Hz west side, which would make clocks like that completely fall over itself.
I've yet to see one clock in Japan which takes the AC frequency as its base. However, large, not battery driven devices like timers for heaters (mechanical timers) have often different scales for 50Hz and 60Hz Those do lock to the AC frequency unfortunately. So if you want to have the living room heated at 7am, you better know whether you live in the 50 or 60Hz part of in Japan.
It's the difference between "freezing" and "feel comfortable".
Better heaters have a built in clock with a built in resonator (whatever type, probably the cheapest available doing the job).
vmware uses files as virtual disks. 2GB would be a really, really small disk.
UML does the same, using the loop device feature of Linux. Again, a filesystem in a file. Again, 2GB is not much. Simulating 20GB would need 10 files.
Feels like 64kbyte segments somehow...and I really don't want to have those back.
Come to think about it, the note is true if you add a line like "...on this computer".
Even nowadays, if you happen to have a CP/M computer, a 10MB hard disk would be fine. (Actually a 1.4MB floppy would be fine.) If you'd put more disk space in it, no one could use it (in a sensible way IMHO). You need much more space to use audio and video, but on such a computer with 8-bit CPU, no audio hardware and a text based screen, who would like this (servers don't count).
I'm going to log it in my firewall, then join the lawsuit, with my logs as proof that I was deceived by these scoundrels.
You want to show them you logged them on your own firewall (probably self installed, running a hard-core version of Linux like Slackware or something similar non-intuitive), and then you fell for their stupid trick of Windows-lookalike silly error messages?
Like this:
Your honour, I tracked down their IP address using tools like nslookup and dig, entered it into my self-installed firewall running Linux using iptables, marking those packets and sending them to metalog (which if I may add, beats sysklogd hands down), so I have a proof that I fell for their ad. Here is my printout of my Gnome desktop (made with xwd and xwud and gimp just because I can). They deceived me! See!
Like this? They have the IBM noisy type keyboards everyones either loves or hates. And they have same-layout-less-noise too. I have no idea whether the latter ones are as good as the noisy ones though.
I have one IBM keyboard. Heavy. Noisy. And I love it.
I was waving the same idea as you have. Since Firewire can support multiple masters (unlike USB) and technically there is no difference in hardware for client and master, it is technically possible to use a PCI hostadapter, given the correct driver software, to emulate a Firewire external disk (or many disks) to another computer. Apple calles it "target mode" for their notebooks.
Now from what I heard and tried, this is not possible given the usual Firewire external disks, as those are "bridges" from IDE to Firewire, not full Firewire controllers. I was not able to see the disk on a second computer connected by the second Firewire port of the disk.
However since you want to emulate the disk using a full blown computer, it's up to the software to do that. Some people have pointed to the Oracle project handling this (or somewhat handling this, as I cannot find out, since their "create a new account" page throws a JSP error...
It might be not exactly what you need, but might be a good starting point.
For myself, I came to the conclusion, that it's not as useful as I thought as first, since you cannot boot from Firewire (Macs can, I know, but my computers are no Macs). And for me a NAS featuring RAID, LVM, xfs/ext3 do support what I mainly want: resizing partitions/space in a flexible way for the client PCs.
But having a SAN on top of Firewire would be nice and would be getting useful as soon as you have bootable Firewire cards. I anyhow wonder why no company does those. You can boot off USB (onboard USB), Ethernet, SCSI, floppy, IDE, but even on on board Firewire controllers you cannot boot from them. This would enable you to have a no-disc PC, running usual OS'es, not just special ones which can boot via NFS (dunno about Win2k).
The AGP bus concept was created to move textures and I guess hardware driver programmers optimized for this.
A quote from the readme file: In the second mode it renders, displays and downloads the same image to the PC.
This is probably not what driver programmers were expecting. Wrong direction of data.
Just an example for a vertical market here in Tokyo:
One of our customers (my employer does IT support) had a "membership" program and a "invoice program" to charge members for services those members use.
Those were well working, but it was written with Novell in mind, the database is file based (read: dBase and Access style), running on DOS (but fortunately running in the DOS box of Win2k). The DBs were independant of each other, so adding a new member involved actions in both databases. Of course it was no longer extendable, had many features missing, was not scaleable. Basically it was out of date and no one could fix it. However, everyone used it as it was the best (and only) they had.
Then the management decided to clean this one up and get new programs and merge them into one.
I don't know how much they payed for it, but the whole conversion is still running after 6 monthes now, they are using 2 new servers running MS SQL Server 2000, all clients need a local Access installation as the query program was written for Access, they have permanently 3 people from that consulting firm in their building, and some more programmers somewhere else, they are using "Great Plains" (dubbed "Great Pain" after a short time running live). Calculate all those numbers and do a guess how much that costs.
We did the member query part using PHP, a web page and a SQL Server connector for PHP (well, actually the Sybase one). Our solution is better, much cheaper, easily extendable by them (if they want, but of course we do this too if they prefer that).
We did nothing about the DB as it was the best running part of the whole project.
And we have more customers running really odd DOS programs (including programs which do not run in a Win2k DOS box).
The point of me showing this example is: there is a vertical market. It's big, it's big money. The problem for newcomers is, how to approach those companies. I would say, it's often luck (being at the correct place at the correct time to offer something they are currently looking for), lots of communication (user must feel comfortable with you, that includes sometimes knows a potential customer for a year without work being done), and of course "word-by-mouth" which is by far the best way (cheap too). All those managers know each other. And if one boasts about the new programs they can use and they got it cheap and the service is great, that will get the attention of ther managers. (Remember: "cheap" for managers is "a whole lot of money" for a startup company of 2 people.)
The nasty part is finding those customers. Once you have a solution for this, go and do good work.
This worked for a friend of mine and me: we got our first customer by knowing an owner of another unrelated company quite well. (Once in a while going out and drinking a beer, fixing small network problems, selling PCs and keep their 1-server-4-PCs network running). And that other company was looking for IT support. Starting here, we got more clients mostly by the managers talking to each others.
I have never seen a toilet which does not include paper as an option. There are many people who do not like this watering.
So you always have the backup in form of old fashioned paper. Like we do in our office: print out
all documentations on paper, so in case the server does not work...but that's a different problem.
Re:PC Weasels are often better than KVM switches.
on
USB KVMs Compared
·
· Score: 1
KVM switches are okay some of the time, but PC Weasels rock!
For remote management (remote=more than 100m) a KVM switch is not an option. The PC Weasel gets around a problem of the consumer PCs, which are not built to be managed remotely. As such, it's clever hard and software. But IMHO it is more cost effective to get a "server board" instead of the usual consumer parts to build a server. They often have a serial console for BIOS access built in (like those from ex-Compaq
or a nice description for another OS here).
This solves the BIOS problem. Does anyone need remote controlability for anything else?
Re:What about video quality over long distances?
on
USB KVMs Compared
·
· Score: 1
The problem is, how do I run video from the downstairs rack to my office (easily a 40' run)?
A friend of mine solved the problem in another way. Instead of having the server room downstairs, he put it next to his working room. Only one set of cables (KVM, and one external SCSI CD-ROM) goes to his desk.
Cable length is about 2m, so no problem at all.
And using a KVM switch, he was omfortably working on 2 computers in a very quiet room.
Of course this is not always possible, but it simplifies the problem a lot, as long as you are flexible about the place of your working room and the server room relative to each other.
In all, 16,000 computers, including machines in every dorm room, will be linked over the coming year to a fiber-optic network that delivers data at up to one gigabit per second.
If I bought100 machines (Florida has probably many more), I could easily take one apart to find out how it works. This would solve the problem, as then the government knows (or should be able to find out) if those machines are ok or not. Trusting one company which says "Our products are fine. You can trust us." without any verification of that claim by independant auditors is just plain stupid.
By checking out one machine, they can lose at most US$3500 if they break it in this process. But they can win confidence for themself and the public in any case, if they break it or not.
Last time I checked, US$3500 was not much for a local government like Florida, so why are they not taking that small risk?
There is also the Macintosh:
Pros: Extremely easy to use, rock solid OS [...]
Cons: Mac OS X still a young OS and there will be bumps in the road. [...]
My wife has a Mac with OS 9.2 and OS X (10.1.4). And while the user interface is useable (far from perfect, but compares very well to Windows and KDE/Gnome), both OS 9 and OS X are not rock solid. And it's not the hardware (e.g. bad RAM)
as we changed to a new Mac lately and OS 9 and OS X freeze once in a while (about twice in a week).
I like OS X (the GUI and how well it all works), but the myth of the stable MacOS is just that.
And while applications could cause this freezing, on OS X thaat should never ever happen.
In essence I can second that. Doing IT support of all kind (from desktop support to setting up large scale WANs) I know customers do those things:
The computer must run all the time, but when they see the price for a full clustered shared RAID system, they suddenly can accept a downtime of 1 day (complete hardware failure).
They want to keep backup data for at least one year, but at the price of those SDLT taped, they tend to choke and cut it down a lot.
Fully redundant links (downtime is not acceptable), but in the end they choose a simple T1 (with service agreement for 99.x% uptime guarantee).
The point of those examples? Customers have wishes without the full knowledge of the consequences. Like the boss who thinks "Speed is everything". While it now is very convinient to say "My boss said so, I know it will fail, but he will be blamed." this does not work. Customers (and bosses) have dreams and wished, but it's up to us (engineers, programmers) to pull them down to Earth and explain them why it's a bad idea and how to compromise, so everyone is happy. So far, this worked very well for me. I bet this works for most bosses too.
PS: I know there are some bosses/customers/etc. who absolutely know better, no matter what you do.
And the answer to the problem is: use look up tables to gain speed if there is memory available, drop them if memory gets tight, and do a graceful abort if memory is full. And limit the valid input to useable amounts by definition (Specs).
Slashdot Mirror
I was working for the IT department of an austrian company in Tokyo. The company grew and office space is expensive, especially if you need to move to another building. The simple fix was, to let some of the developers and translators, who were working already several years at the company, work from home. After all, those do not need physical appearance and they prefer (due to the nature of work) a quiet environment. Something which is difficult to get in a japanese company.
Worked out well, as it was easy to check they are working by checking the results. The employees (not all wanted to work from home) were generally happy, some office space was saved, travel money (paid usually by the company) was saved, in the end, everyone was happy.
I think the trick in this excercise was, so let experienced workers work from home. People who are known to be able to motivate themself. And as everyone could check the productivity, the usual problem of teleworking, not being able to tell if the employee watches TV or works 8 hours, did not apply here.
Life gets more complicated when coming originally from Germany, where VAT is always included in the advertized price (for end-users, not for businesses) and going to Japan, where 5% tax is mostly added. About 90% of all times I need to add 5% at the cashier. In the other cases I don't need to. Now that makes calculation the price to pay complicated.
My experience reflects this too. While theoretically and actually practically (when doing large file transfers) the speed is enough, you get dropped frames as the data does not flow as a steady flow. Even a modest 1.5MBit/s divx is far from perfect when watching it being transfered from a (fast) file server to a (fast) computer. It's ok for short clips, but watching a movie is out of question. I prefer to copy the movie to disk and then watch it. Yet, I can move about 600kByte/s data on the 11MBit/s wireless LAN.
Wouldn't that be even better? This way they know what to expect from you and (hopefully) that will result in less spam from them. Simple calculation: this potential customer costs us more maney than we can possible earn from this him/her.
And since you do nothing illegal, no need for fear! In the easiest case send them their empty envelope back. You can always say that you forgot to put something in.
It's a nice way to check a connection is not made, that packets do not go out of one or another interface, that traffic is encrypted. tcpdump can do the same (except follow TCP traffic, which is very enlightning for users who like telnet).
So while Ethereal does not increase security by itself, it does add security by making it possible to check out the packets. That makes is IMHO a security tool.
Here in Tokyo, the place where you want to be when you want Internet connections for cheap, the standard DSL service is 12Mbit/s down, 1Mbit/s up. For abour Yen 3000 (about US$ 26). And so far no restrictions. And it's fast (within Japan 900kbyte/s if the server is fast enough, to USA usually 200kbyte/s).
Everything else in Japan and especially in Tokyo is expensive. But Internet is as cheap as you can imagine.
Starting at Ultra320, Quick Arbitration is now available.
Given that this is cutting edge of parallel SCSI, I can understand Sun to skip anything older. However I yet have to see a significant performance gain from going from U2W (80 MB/s) to FC. Arbitriation might be slower on U2W, but FC contains routing informations in each packet, which parallel SCSI lacks.
Personally, I am very happy with the good old parallel SCSI. Even cutting edge drives like Cheetah 15k.3 are really fast, even when I cannot push them to their limit.
Houses that are supposed stay cool in the summer should be well insulated.
That's what I thought and that's what common sense is. But not here in Japan. Summer is humid and long (not just 1 week humid hot like it is usually in Germany once in a year). 3 Monthes humidty of easily 70% up to 90% and temperatures of 30+ centigrades. At night it often stays between 25 and 30 centigrades.
Now imagine a cool inside...you can nearly feel the mold growing. It would be ok if you never need to leave the house, but open the door and, baaam, humidity comes in.
So the perfect solutions would be something that keeps the inside insulated from the outside (temperature wise), yet allows a free flow of air. That seems to be tricky. I have not found a solution.
Old houses in Japan have a clever air circulation system: the top of the room walls are open (can be closed in winter) and as hot air moves up, it's pushed out by a slight breeze. Many of the outside walls can be opened using sliding doors. So if there is any breeze outside, it's not too hot inside. 30 degrees outside or not.
Unfortunately it became unpopular to have the top part of room walls open, as there is no privacy any more. Any noise in a house is audible in all connected rooms. And air conditioners completely changed the way of cooling, introducing the humidity problem. And all this is amplified by having a lot of concrete, increasing the outside temperatures even more on a sunny day. Most days in summer are sunny. That's modern Tokyo. But that's another problem.
Modern houses in Japan are really built to last 20 years. Maybe 30. I live in one which is 16 years old and the owner probably would like to rebuilt it (which is not soo expensive as one might think, as the ground is a very expensive part of a house, and the house itself is built by "wood and paper and plastic". Not many stones here...
Any idea why there are 0 of them in Japan? Japan runs on 50Hz east side and 60Hz west side, which would make clocks like that completely fall over itself.
I've yet to see one clock in Japan which takes the AC frequency as its base. However, large, not battery driven devices like timers for heaters (mechanical timers) have often different scales for 50Hz and 60Hz Those do lock to the AC frequency unfortunately. So if you want to have the living room heated at 7am, you better know whether you live in the 50 or 60Hz part of in Japan. It's the difference between "freezing" and "feel comfortable".
Better heaters have a built in clock with a built in resonator (whatever type, probably the cheapest available doing the job).
vmware uses files as virtual disks. 2GB would be a really, really small disk. UML does the same, using the loop device feature of Linux. Again, a filesystem in a file. Again, 2GB is not much. Simulating 20GB would need 10 files.
Feels like 64kbyte segments somehow...and I really don't want to have those back.
Out of curiousity: how do you get rid of the system drive letter (C: usually)?
Come to think about it, the note is true if you add a line like "...on this computer".
Even nowadays, if you happen to have a CP/M computer, a 10MB hard disk would be fine. (Actually a 1.4MB floppy would be fine.) If you'd put more disk space in it, no one could use it (in a sensible way IMHO). You need much more space to use audio and video, but on such a computer with 8-bit CPU, no audio hardware and a text based screen, who would like this (servers don't count).
I'm going to log it in my firewall, then join the lawsuit, with my logs as proof that I was deceived by these scoundrels.
You want to show them you logged them on your own firewall (probably self installed, running a hard-core version of Linux like Slackware or something similar non-intuitive), and then you fell for their stupid trick of Windows-lookalike silly error messages?
Like this:
Your honour, I tracked down their IP address using tools like nslookup and dig, entered it into my self-installed firewall running Linux using iptables, marking those packets and sending them to metalog (which if I may add, beats sysklogd hands down), so I have a proof that I fell for their ad. Here is my printout of my Gnome desktop (made with xwd and xwud and gimp just because I can). They deceived me! See!Like this? They have the IBM noisy type keyboards everyones either loves or hates. And they have same-layout-less-noise too. I have no idea whether the latter ones are as good as the noisy ones though.
I have one IBM keyboard. Heavy. Noisy. And I love it.
Hi,
I was waving the same idea as you have. Since Firewire can support multiple masters (unlike USB) and technically there is no difference in hardware for client and master, it is technically possible to use a PCI hostadapter, given the correct driver software, to emulate a Firewire external disk (or many disks) to another computer. Apple calles it "target mode" for their notebooks.
Now from what I heard and tried, this is not possible given the usual Firewire external disks, as those are "bridges" from IDE to Firewire, not full Firewire controllers. I was not able to see the disk on a second computer connected by the second Firewire port of the disk.
However since you want to emulate the disk using a full blown computer, it's up to the software to do that. Some people have pointed to the Oracle project handling this (or somewhat handling this, as I cannot find out, since their "create a new account" page throws a JSP error... It might be not exactly what you need, but might be a good starting point.
For myself, I came to the conclusion, that it's not as useful as I thought as first, since you cannot boot from Firewire (Macs can, I know, but my computers are no Macs). And for me a NAS featuring RAID, LVM, xfs/ext3 do support what I mainly want: resizing partitions/space in a flexible way for the client PCs.
But having a SAN on top of Firewire would be nice and would be getting useful as soon as you have bootable Firewire cards. I anyhow wonder why no company does those. You can boot off USB (onboard USB), Ethernet, SCSI, floppy, IDE, but even on on board Firewire controllers you cannot boot from them. This would enable you to have a no-disc PC, running usual OS'es, not just special ones which can boot via NFS (dunno about Win2k).
Harald
Something tells me he's not talking about IDE drives if he's building an array of 12 of them.
Look here. 12 IDE disks on one RAID controller.
The AGP bus concept was created to move textures and I guess hardware driver programmers optimized for this.
A quote from the readme file: In the second mode it renders, displays and downloads the same image to the PC.
This is probably not what driver programmers were expecting. Wrong direction of data.
Just an example for a vertical market here in Tokyo:
One of our customers (my employer does IT support) had a "membership" program and a "invoice program" to charge members for services those members use. Those were well working, but it was written with Novell in mind, the database is file based (read: dBase and Access style), running on DOS (but fortunately running in the DOS box of Win2k). The DBs were independant of each other, so adding a new member involved actions in both databases. Of course it was no longer extendable, had many features missing, was not scaleable. Basically it was out of date and no one could fix it. However, everyone used it as it was the best (and only) they had.
Then the management decided to clean this one up and get new programs and merge them into one.
I don't know how much they payed for it, but the whole conversion is still running after 6 monthes now, they are using 2 new servers running MS SQL Server 2000, all clients need a local Access installation as the query program was written for Access, they have permanently 3 people from that consulting firm in their building, and some more programmers somewhere else, they are using "Great Plains" (dubbed "Great Pain" after a short time running live). Calculate all those numbers and do a guess how much that costs.
We did the member query part using PHP, a web page and a SQL Server connector for PHP (well, actually the Sybase one). Our solution is better, much cheaper, easily extendable by them (if they want, but of course we do this too if they prefer that). We did nothing about the DB as it was the best running part of the whole project.
And we have more customers running really odd DOS programs (including programs which do not run in a Win2k DOS box).
The point of me showing this example is: there is a vertical market. It's big, it's big money. The problem for newcomers is, how to approach those companies. I would say, it's often luck (being at the correct place at the correct time to offer something they are currently looking for), lots of communication (user must feel comfortable with you, that includes sometimes knows a potential customer for a year without work being done), and of course "word-by-mouth" which is by far the best way (cheap too). All those managers know each other. And if one boasts about the new programs they can use and they got it cheap and the service is great, that will get the attention of ther managers. (Remember: "cheap" for managers is "a whole lot of money" for a startup company of 2 people.)
The nasty part is finding those customers. Once you have a solution for this, go and do good work. This worked for a friend of mine and me: we got our first customer by knowing an owner of another unrelated company quite well. (Once in a while going out and drinking a beer, fixing small network problems, selling PCs and keep their 1-server-4-PCs network running). And that other company was looking for IT support. Starting here, we got more clients mostly by the managers talking to each others.
I have never seen a toilet which does not include paper as an option. There are many people who do not like this watering.
So you always have the backup in form of old fashioned paper. Like we do in our office: print out all documentations on paper, so in case the server does not work...but that's a different problem.
KVM switches are okay some of the time, but PC Weasels rock!
For remote management (remote=more than 100m) a KVM switch is not an option. The PC Weasel gets around a problem of the consumer PCs, which are not built to be managed remotely. As such, it's clever hard and software.
But IMHO it is more cost effective to get a "server board" instead of the usual consumer parts to build a server. They often have a serial console for BIOS access built in (like those from ex-Compaq or a nice description for another OS here).
This solves the BIOS problem. Does anyone need remote controlability for anything else?
The problem is, how do I run video from the downstairs rack to my office (easily a 40' run)?
A friend of mine solved the problem in another way. Instead of having the server room downstairs, he put it next to his working room. Only one set of cables (KVM, and one external SCSI CD-ROM) goes to his desk. Cable length is about 2m, so no problem at all. And using a KVM switch, he was omfortably working on 2 computers in a very quiet room.
Of course this is not always possible, but it simplifies the problem a lot, as long as you are flexible about the place of your working room and the server room relative to each other.
No you don't:
In all, 16,000 computers, including machines in every dorm room, will be linked over the coming year to a fiber-optic network that delivers data at up to one gigabit per second.
If I bought100 machines (Florida has probably many more), I could easily take one apart to find out how it works. This would solve the problem, as then the government knows (or should be able to find out) if those machines are ok or not. Trusting one company which says "Our products are fine. You can trust us." without any verification of that claim by independant auditors is just plain stupid.
By checking out one machine, they can lose at most US$3500 if they break it in this process. But they can win confidence for themself and the public in any case, if they break it or not. Last time I checked, US$3500 was not much for a local government like Florida, so why are they not taking that small risk?
There is also the Macintosh:
Pros: Extremely easy to use, rock solid OS [...]
Cons: Mac OS X still a young OS and there will be bumps in the road. [...]
My wife has a Mac with OS 9.2 and OS X (10.1.4). And while the user interface is useable (far from perfect, but compares very well to Windows and KDE/Gnome), both OS 9 and OS X are not rock solid. And it's not the hardware (e.g. bad RAM) as we changed to a new Mac lately and OS 9 and OS X freeze once in a while (about twice in a week).
I like OS X (the GUI and how well it all works), but the myth of the stable MacOS is just that. And while applications could cause this freezing, on OS X thaat should never ever happen.
In essence I can second that. Doing IT support of all kind (from desktop support to setting up large scale WANs) I know customers do those things:
The computer must run all the time, but when they see the price for a full clustered shared RAID system, they suddenly can accept a downtime of 1 day (complete hardware failure).
They want to keep backup data for at least one year, but at the price of those SDLT taped, they tend to choke and cut it down a lot.
Fully redundant links (downtime is not acceptable), but in the end they choose a simple T1 (with service agreement for 99.x% uptime guarantee).
The point of those examples? Customers have wishes without the full knowledge of the consequences. Like the boss who thinks "Speed is everything". While it now is very convinient to say "My boss said so, I know it will fail, but he will be blamed." this does not work. Customers (and bosses) have dreams and wished, but it's up to us (engineers, programmers) to pull them down to Earth and explain them why it's a bad idea and how to compromise, so everyone is happy. So far, this worked very well for me. I bet this works for most bosses too.
PS: I know there are some bosses/customers/etc. who absolutely know better, no matter what you do. And the answer to the problem is: use look up tables to gain speed if there is memory available, drop them if memory gets tight, and do a graceful abort if memory is full. And limit the valid input to useable amounts by definition (Specs).