This is somewhat off-topic, but I go to the college at which Peter Suber teaches. I have yet to take a course from him, but I have read many of his writings. His website has interesting stuff not only on copyright law but also on computer science and philosophy.
Actually, it's closer to 170, which isn't much higher than with all the services off (~130). And, while this isn't a front-line server, the processes are hardly idle either.
A lot of internships aren't paid up front, but being able to put it on a resume and get references from it can definitely make an internship a wise investment.
If you're using GNOME or KDE then, yes, I have to agree with you. But my dual 450MHz Xeon Linux machine with 256MB of RAM still blows the socks off those 2.5GHz monsters running XP. I use XFce as a window manager, and couldn't be happier. I run Apache (mod_perl, mod_php, and mod_ssl), ProFTPd, MySQL, NFS, NIS, OpenLDAP, Samba, Sendmail, and UW-IMAP, and there's still enough juice for it to function well as a workstation.
Except it's PDF that actually has the encryption support built-in. Breaking that encryption is a violation of the DMCA. The document that the bill was in was actually protected by a 40-bit key. In this case, it just prevented modification, but it could just as easily prevent anything else. PostScript, OTOH, is just a page layout language, and has none of the nasty strings coming from the DMCA.
Does anyone else find irony in the fact that the bill is distributed in Adobe PDF format? I seem to recall them bringing a certain Sklyarov character running afoul of them and the DMCA a few years ago....
The government needs to at least provide the option of PostScript, text-only or other non-corporate formats for bills and laws.
My apologies to those who have the time to calculate that out. My point still stands, though.
How can they say that, after not even a year of development, a product has 99.995% availablity? IBM puts years of assessment into their MVS, z/OS, OS/360, OS/390, and OS/400 mainframe operating systems to get their 5+ nines of availability.
Windows Server 2003 hit 99.995 percent availability at the Release Candidate 1 (RC1) stage last summer
Correct me if I'm wrong, but doesn't this mean that they must have had something running a year before that? 4.5 9's of availability is what, a day of down-time per year?
If you can't trust your co-lo provider, I suggest you move pretty damn fast. Ask them if they abide by the SAGE Code of Ethics, and even see if you can get them to sign it as part of the contract.
Why bother? Nothing, not even encryption, beats physical security. Put a nice strong lock (keyed, number pad, biometric, whatever) on the server room door, give out access on a need-to-have basis, and you've bought more security than encryption can get you, without any of the performance hits.
Apple should ship a pair of earplugs with any workstation xServes then. My school has a 16-node cluster of dual 1GHz xServes in a rack (sans RAID box), and it sounds like a wind tunnel. It's even louder than the old 16-node cluster of dual 1GHz 2U VA Linux boxen, and they were the reason that the lab they are in is no longer used; it's too loud to hold any kind of conversation or discussion with even one cluster in the room. A few of the spare VA Linux machines are used as workstations, and they're almost too loud when they're strapped underneath tables.
From what I have seen, not having the IT staff train all the employees is a good to avoid friction between IT and the other departments. Have IT train a few enthusiastic and knowledgable non-IT people, and then have those people go out and train the rest of the employees. The teachers will be able to sympathize more with regular employees, and the regular employees will look at the teachers with more respect than if they were IT.
In my computing experience, I have used NetBSD, FreeBSD, and OpenBSD. I can't stand FreeBSD's installation process. I always end up going in a loop with assigning disks or labels. The ports collection works most of the time, but is damn annoying when it doesn't. OpenBSD's installation is fine, but post-install it's a pain getting everything up to the point of functionality. NetBSD strikes a good balance between the two, with the purely text-based installer but also a functional post-install.
I am also happy knowing that all my non-SMP servers can be running the same OS, no matter if they are Intel, MIPS, SPARC, Alpha, VAX, or otherwise. Hardware age doesn't matter, either. I use an old 486DX with 32MB of RAM as a DNS/DHCP/SMTP/IMAP/LDAP server for a network, and it handles the load beautifully.
I remember I once had an errant Perl script that kept on spawning itself, sucked all the available memory, and pegged the CPU load average to around 70. It was slow, but I managed to log in, kill all the Perl scripts, and everything returned to normal instantly. No crash, no fallout. I'd like to see Windoze 2000 (or even Linux) do that!
As another anecdote from my work at my school, I have a professor who, while on our FreeBSD 4.6.2 server, forgot to grep for nfsd in an awk script when he was trying to kill and restart NFS. This killed all the processes on the server, and left it without init, basically in an unusable state. I can recall doing that in NetBSD, and it appears that there is a kernel function that checks that init is still running and will respawn it if it dies, and then place the machine in single-user mode. I would still have had to run over to the science building and get things going again, but our uptime would have been preserved, and I would have been able to do administrative work immediately.
It's a pity that NetBSD doesn't have more users than it does. It has got to be one of the most capable open-source OSs, and I certainly prefer it over some commercial UNIXs such as AIX.
I think you are confusing legality and morality. At least in my world, killing is something that should be avoided at any cost. While Saddam might legally be responsible for putting SAMs and AAAs in civilian areas, it is the United States that bears the moral responsibility if the civilians are killed. It doesn't matter why we're there, although in this case I would say the United States' intentions are far less than noble.
One has to keep in mind that these programs are not designed to work with journaling filesystems. The presence of a journal means that the wiper is not actually certain that the blocks were overwritten.
If one is giving away a hard drive, it is a good idea to low-level format the drive (if it is SCSI), and/or create a partition spanning the entire disk, dd from/dev/zero on to the disk until it is full, and then use a wiper to delete that file.
I started using REXX in 1990 and it was my primary scripting language until I discovered real [perl.org] scripting [gnu.org]
languages [python.org].
REXX can do many of things that Perl and other commonly-used languages can do. If you don't like it, C, Perl, Python, Tk, and a host of other languages are available for OS/2. REXX is by no means the only available language.
OS/2's command line is no more powerful than the DOS command line. It pales in comparison to the UNIX shell, which is why several companies released enhanced shells for for OS/2.
If you don't like OS/2's command line, then you can run bash instead. Other Unix shells, including ksh and csh have been ported as well. This, of course, gives you access to Unix shell scripting as well, providing another replacement for REXX scripting.
I developed for OS/2 over the course of 12 years at a factory with hundreds of OS/2 workstations. The stability of later versions of the OS/2 kernel is impressive: I've seen the kernel keep chugging along after the desktop hangs on a number of occassions. But what good is that when other layers of the system are so confounded that the only thing that solves the problem is a reboot?
The Workplace Shell can be restarted easily without rebooting, thereby solving the problem of a hung desktop.
So best of luck in your advocacy of a dying OS (and in the improvement of your manners) but I stand by my statement: OS/2 is not a good server operating system.
Then I would expect that banks would not use it quite so much. Banks nowadays require 4+ nines of uptime, which seems to imply to me that OS/2 is one of the most capable server OSs in existence.
6. Items that were rarities due to supply or marketing decisions (I bet that 1.2GHz Hammers they sent around for demonstration will be worth something, and did anyone ever get a 160MHz Am5x86?)
No, but I have the 133MHz version of that chip. It was used to upgrade a 33MHz i486 about seven years ago. At the time, the box was running OS/2, but now the OS/2 installation has been moved to a faster machine. After removing OS/2, I installed NetBSD on it, and now use it as a DHCP/DNS/IMAP/SMTP server for my LAN. The thing gets fabulous uptime, and only goes down for power outages that last longer than the UPS can hold out . They just don't make 'em like they used to....
Is the MSI KT4V-L the original nVidia nForce? If it is, I have that too, and am also running Win2k (SP3) on it. I have the onboard NIC running at between 2.5 and 3.5MB/s. Not great, but good enough for my purposes.
I have this card on a PII-350MHz Red Hat Linux machine, and an integrated one in a 1.3GHz Celeron laptop dual-booting Windoze 2000 and Debian Linux. In Linux, I never get above 5MB/s on a 100baseT-FX link, but in Windoze the limit is around 1.5MB/s. I guess Windoze has even worse support for this chipset than Linux.
It's kind of nice having something that doesn't weigh over 20kg and suck over 100W of electricity just to display CPU stats. There's lots of use for this in server rooms, where such information can be displayed without having to fire up a monitor and switch the KVM over.
In the early 1980's, "Windows represented a new kind of product for personal computers," Mr. Gates wrote. "It was not part of any existing product category."
IIRC, both Xerox and Apple had windowing systems at that point. Bull Gates seems to have a long history of spewing bullshit.
You're worried about 18-meter 4-ton helium filled balloons over heavily-populated areas? We already have 80-meter 40-ton airplanes flying over heavily-populated areas, and they're filled with highly flammable jet fuel to boot.
Slashdot Mirror
This is somewhat off-topic, but I go to the college at which Peter Suber teaches. I have yet to take a course from him, but I have read many of his writings. His website has interesting stuff not only on copyright law but also on computer science and philosophy.
Actually, it's closer to 170, which isn't much higher than with all the services off (~130). And, while this isn't a front-line server, the processes are hardly idle either.
A lot of internships aren't paid up front, but being able to put it on a resume and get references from it can definitely make an internship a wise investment.
If you're using GNOME or KDE then, yes, I have to agree with you. But my dual 450MHz Xeon Linux machine with 256MB of RAM still blows the socks off those 2.5GHz monsters running XP. I use XFce as a window manager, and couldn't be happier. I run Apache (mod_perl, mod_php, and mod_ssl), ProFTPd, MySQL, NFS, NIS, OpenLDAP, Samba, Sendmail, and UW-IMAP, and there's still enough juice for it to function well as a workstation.
Except it's PDF that actually has the encryption support built-in. Breaking that encryption is a violation of the DMCA. The document that the bill was in was actually protected by a 40-bit key. In this case, it just prevented modification, but it could just as easily prevent anything else. PostScript, OTOH, is just a page layout language, and has none of the nasty strings coming from the DMCA.
Does anyone else find irony in the fact that the bill is distributed in Adobe PDF format? I seem to recall them bringing a certain Sklyarov character running afoul of them and the DMCA a few years ago....
The government needs to at least provide the option of PostScript, text-only or other non-corporate formats for bills and laws.
My apologies to those who have the time to calculate that out. My point still stands, though.
How can they say that, after not even a year of development, a product has 99.995% availablity? IBM puts years of assessment into their MVS, z/OS, OS/360, OS/390, and OS/400 mainframe operating systems to get their 5+ nines of availability.
I would imagine a prime reason would be that Yellow-Dog has less fluff and eye-candy than OS X, which would allow more efficient resource use.
I would be curious to know whether the person got power management working on it.
If you can't trust your co-lo provider, I suggest you move pretty damn fast. Ask them if they abide by the SAGE Code of Ethics, and even see if you can get them to sign it as part of the contract.
Why bother? Nothing, not even encryption, beats physical security. Put a nice strong lock (keyed, number pad, biometric, whatever) on the server room door, give out access on a need-to-have basis, and you've bought more security than encryption can get you, without any of the performance hits.
Apple should ship a pair of earplugs with any workstation xServes then. My school has a 16-node cluster of dual 1GHz xServes in a rack (sans RAID box), and it sounds like a wind tunnel. It's even louder than the old 16-node cluster of dual 1GHz 2U VA Linux boxen, and they were the reason that the lab they are in is no longer used; it's too loud to hold any kind of conversation or discussion with even one cluster in the room. A few of the spare VA Linux machines are used as workstations, and they're almost too loud when they're strapped underneath tables.
This is the second story with no comments posted. What's up?
From what I have seen, not having the IT staff train all the employees is a good to avoid friction between IT and the other departments. Have IT train a few enthusiastic and knowledgable non-IT people, and then have those people go out and train the rest of the employees. The teachers will be able to sympathize more with regular employees, and the regular employees will look at the teachers with more respect than if they were IT.
In my computing experience, I have used NetBSD, FreeBSD, and OpenBSD. I can't stand FreeBSD's installation process. I always end up going in a loop with assigning disks or labels. The ports collection works most of the time, but is damn annoying when it doesn't. OpenBSD's installation is fine, but post-install it's a pain getting everything up to the point of functionality. NetBSD strikes a good balance between the two, with the purely text-based installer but also a functional post-install.
I am also happy knowing that all my non-SMP servers can be running the same OS, no matter if they are Intel, MIPS, SPARC, Alpha, VAX, or otherwise. Hardware age doesn't matter, either. I use an old 486DX with 32MB of RAM as a DNS/DHCP/SMTP/IMAP/LDAP server for a network, and it handles the load beautifully.
I remember I once had an errant Perl script that kept on spawning itself, sucked all the available memory, and pegged the CPU load average to around 70. It was slow, but I managed to log in, kill all the Perl scripts, and everything returned to normal instantly. No crash, no fallout. I'd like to see Windoze 2000 (or even Linux) do that!
As another anecdote from my work at my school, I have a professor who, while on our FreeBSD 4.6.2 server, forgot to grep for nfsd in an awk script when he was trying to kill and restart NFS. This killed all the processes on the server, and left it without init, basically in an unusable state. I can recall doing that in NetBSD, and it appears that there is a kernel function that checks that init is still running and will respawn it if it dies, and then place the machine in single-user mode. I would still have had to run over to the science building and get things going again, but our uptime would have been preserved, and I would have been able to do administrative work immediately.
It's a pity that NetBSD doesn't have more users than it does. It has got to be one of the most capable open-source OSs, and I certainly prefer it over some commercial UNIXs such as AIX.
I think you are confusing legality and morality. At least in my world, killing is something that should be avoided at any cost. While Saddam might legally be responsible for putting SAMs and AAAs in civilian areas, it is the United States that bears the moral responsibility if the civilians are killed. It doesn't matter why we're there, although in this case I would say the United States' intentions are far less than noble.
One has to keep in mind that these programs are not designed to work with journaling filesystems. The presence of a journal means that the wiper is not actually certain that the blocks were overwritten.
If one is giving away a hard drive, it is a good idea to low-level format the drive (if it is SCSI), and/or create a partition spanning the entire disk, dd from /dev/zero on to the disk until it is full, and then use a wiper to delete that file.
This BBC article gives a much better answer to the problem faced by banna growers: Bananas could split for good
The gist of the article is, without genetic modifications, bannas might go extinct within the decade.
No, but I have the 133MHz version of that chip. It was used to upgrade a 33MHz i486 about seven years ago. At the time, the box was running OS/2, but now the OS/2 installation has been moved to a faster machine. After removing OS/2, I installed NetBSD on it, and now use it as a DHCP/DNS/IMAP/SMTP server for my LAN. The thing gets fabulous uptime, and only goes down for power outages that last longer than the UPS can hold out . They just don't make 'em like they used to....
Is the MSI KT4V-L the original nVidia nForce? If it is, I have that too, and am also running Win2k (SP3) on it. I have the onboard NIC running at between 2.5 and 3.5MB/s. Not great, but good enough for my purposes.
BTW, did you download the unified driver package from nVidia? That's what I have installed, so you might try that. URL is http://download.nvidia.com/Windows/nForce/2.00/nFo rce_win2k_2.00.exe
I have this card on a PII-350MHz Red Hat Linux machine, and an integrated one in a 1.3GHz Celeron laptop dual-booting Windoze 2000 and Debian Linux. In Linux, I never get above 5MB/s on a 100baseT-FX link, but in Windoze the limit is around 1.5MB/s. I guess Windoze has even worse support for this chipset than Linux.
It's kind of nice having something that doesn't weigh over 20kg and suck over 100W of electricity just to display CPU stats. There's lots of use for this in server rooms, where such information can be displayed without having to fire up a monitor and switch the KVM over.
IIRC, both Xerox and Apple had windowing systems at that point. Bull Gates seems to have a long history of spewing bullshit.
You're worried about 18-meter 4-ton helium filled balloons over heavily-populated areas? We already have 80-meter 40-ton airplanes flying over heavily-populated areas, and they're filled with highly flammable jet fuel to boot.