I would love to take your friend up on it, since we have had some difficulty finding system admins. However, my company is not authorizing outside hires atm.
Isn't this essentially what ethernet does? Two channels broadcast on at the same time, and nothing responds, each one sets a pseudo random time to rebroadcast...the internet runs on this system and doesn't die - why can't radio be set up along similar lines? (heck we used to use repeaters for our radios back in my army days).
Additionally, we are doing wave division multiplexing with T1 circuits, as well as fiber optics using multi spectrum...not to mention DSL across a 2 wire circuit layered on top of voice...
Given all of that, I think it is feasible to build technology to make the concepts discussed in the article reality.
I work in the 'real world' and here is the experience we are looking for:
Linux/Unix - both system programming and system administration experience. Show us how you would automate various features, and integrate different systems together to get real work done quickly. VOIP and Telephony - convergence is not just a buzz word. Java/CGI/XML - web enabled application development is a must. No one I am talking to is considering.NET A plus is experience using Perl/Tk, TCL/Tk, C++ (gcc), emacs, vi, awk, sed, and shell scripting.
Things that will not get you hired:
Primary Microsoft experience; Microsoft certifications mean nothing in our space. I've lost count of how many microsofties come in looking for work, and are totally lost in the datacentre. MBA - you would be surprised at how many folks think 'system administrator' means 'managing people'; if you don't have a technical background, forget it.
I think some folks have a misapprehension that a CS degree only equates to being a programer.
Disabuse yourself of that fallacy, and you will be set.
Case in point: Me.
I started out of college NOT AS A PROGRAMMER - but as a system administrator (I parlayed my brief experience with Unix in college into the job). Once in the position, I automated all of the tedious stuff, then set my sights on some of the other things around me. For example, the technical support folks were using paper tickets; a little database magic and some cgi scripts, and 'Voila!' - an automated ticketing system.
After that, I extended my knowledge - I learned other operating systems the company was using, and I also learned new programming languages. I also started developing my philosophy regarding programming paradigms for large projects by reading everything I could get my hands on and watching what went on with internal projects. Most importantly, I learned how to make disparate systems work together (systems integration).
Finally I got into a major project as a programer and applied and refined the lessons I had learned. Having administered operating systems and databases, I had a unique perspective many of my peers did not have. I volunteered for the difficult tasks, because I knew I would learn more by stretching myself than by sitting back and just coding by rote.
Pretty soon, I was project lead, and then a full blown senior developer. At this stage of the game I deal with technology and implementation issues at a high level; I write specifications, and either implement it (if its trivial) or oversee a vendor or internal team perform the implementation.
That is how a CS major can take you where you want to go; don't limit yourself by setting your expectations too high, or conversely, setting them too low. The biggest key is to just make yourself as valuable as possible to your company and you will get where you are going.
This is probably just a means for the government to collect a list of 'dissidents'.
According to government statistics, there is a direct correlation between gun ownership, people who believe in 'real' freedom (not the PC fakey kind), and anti-telemarking activists.
Can anyone comment on if and how the SCO lawsuit may or may not effect linux kernel distros now or in the future?
I wouldn't throw away the earlier kernels just yet. It may not effect your average mom and pop operation, but a legal injunction would curtail alot of corporate projects that are currently using linux.
It would be nice to see someone respond who is very familiar with the kernel development and these issues.
That is a false stereotype. I have many qualities that would be considered 'computer geek' - whatever that means - and can communicate quite well, thank you.
I prefer to be considered a unique human being, as opposed to being labeled and filed into a neat little cubbyhole. Hopefully your sense of self worth makes you feel the same way. However, I could be wrong about that; you may be perfectly happy assuming the one dimensional attributes of the labels you embrace.
What scares me more than the government misusing its power, is the people here who say "its too difficult - I would give my key away if a gun was to my head..."
If you are an American, you should be ashamed of yourself. This is exactly the situation we faced 200+ years ago. What if the people who founded this nation decided the sacrifice was too hard? Would they have turned turtle and spilled the beans, giving their friends and compatriots a death sentence?
While most moments of your life are spent on less than stellar events, when the time comes to be measured I only hope you will rise to the occaision, instead of slinking away like a dog.
Re:The really interesting thing about dupes
on
IETF to Look at Spam
·
· Score: 1
What they need is some software that can take care of this for them as they are going through the stories; just rewrite SLASH to take care of it (I would, but I am up to my eyeballs in greased potbelly pigs atm).
Why so touchy on the subject of 'Microsoft Bashing'? The ability to thumb our noses at monopolies and bad government is an american pass-time (there is plenty of prior art if you look in the history books).
Could it be you own large amounts of Microsoft stock? Do you work at the company? Or, are you just ignorant of the findings in the antitrust case?
Feel free to bash the open source community, Apple Computing, Sun, HP, IBM, or anyone else you find loathsome. Don't tell me how to regulate my expression (only CowboyNeal and the gang can do that here).
"...who could get past the wall of flaming cow dung..."
Now that describes the situation precisely.
We are having trouble now getting a system administrator for a project because HR classifies 'system administrator' as a low level 'mainframe operator' kind of job, and thus does not allow a proper paygrade for someone who is going to be managing multiple OSes and large numbers of machines with a high need to understand and manage system admin automation.
We need a superstar sysadmin; we will probably get some kid out of college who can't find his butt with both hands.
Do you study the makeup and practices of the development team as part of your analysis? Would you find it useful to know if a team favored one lifecycle methodology over another - and are there any correlations you have seen along these lines?
Stories should not be altered and then attributed to a particular/. nick. If a 'mixed' approach is needed, then the editor should write the piece, giving credit for all of the 'pieces' where credit is due.
I think his point is there is a perception that because our fingers aren't constantly tapping on the keyboard, we are not doing any work.
The reality is Unix/Linux system administrators are not paid what they are really worth to the organization because the suits don't understand all of the things we do on a constant basis to:
1. Develop software tools to automate various activities and save man hours for the business. Good design takes time and thought, things not available in the factory line and cube farm mentality of most organizations.
2. Mitigate disaster (natural, software [stupid programmers/viruses], or otherwise) by designing systems and networks that work - and if they fail recover automagically - or with minimal time and effort.
I consistently have the boss come by and say, "I don't see you working", or, "what are you no-goods up to?". There is no conception of the genius that is put into the systems, and its value to the company. System administrators are not interchangeable; good ones are few and far between.
Elbereth - your tirade does not become you. If you don't grok unix/linux/bsd, or don't care for our inside jokes, kindly keep it to yourself.
Xerithane - Linux is Posix compliant, just as BSD and the original AT&T System V. Per http://www.cs.buffalo.edu/~smgallo/SEpaper/node14. html: "In 1968, AT&T's Bell Labs began development on the Unix operating system. Shortly thereafter, many hardware vendors had developed their own flavor of Unix for their particular machines, which made porting software between platforms difficult. Recognizing the need for a standard to make porting applications between operating systems simpler, the IEEE developed POSIX in 1988 and in 1990 it became International Standard ISO/IEC 9945-1:1990 [14]. The United States Government has adopted POSIX as a Federal Information Processing Standard (FIPS 151) inspiring system vendors to do the same. The list of vendors that have announced support for POSIX includes AT&T, Control Data Corporation, Digital Equipment Corporation, Grumman Data Systems, Hewlett-Packard, Intel, IBM, Microsoft, and Sun Microsystems.
POSIX defines a standard way for an application program to obtain basic services from the operating system. It describes a set of functions derived from AT&T UNIX (System V) and Berkeley Standard Distribution UNIX (BSD). All POSIX-conforming systems must implement these functions, and programs that follow the POSIX standard use only these functions to obtain services from the operating system and the underlying hardware [14].
Although POSIX is based on AT&T and BSD UNIX, it is not an operating system itself. It simply defines an interface between software and their libraries. This allows non-UNIX operating systems to conform to the POSIX standard."
And from http://search390.techtarget.com/sDefinition/0,,sid 10_gci214309,00.html more details: "Informally, each standard in the POSIX set is defined by a decimal following the POSIX. Thus, POSIX.1 is the standard for an application program interface in the C language. POSIX.2 is the standard shell and utility interface (that is to say, the user's command interface with the operating system). These are the main two interfaces, but additional interfaces, such as POSIX.4 for thread management, have been developed or are being developed. The POSIX interfaces were developed under the auspices of the Institute of Electrical and Electronics Engineers (IEEE).
POSIX.1 and POSIX.2 interfaces are included in a somewhat larger interface known as the X/Open Programming Guide 4.2 (also known as the "Single UNIX Specification" and "UNIX 95"). The Open Group, an industry standards group, owns the UNIX trademark and can thus "brand" operating systems that conform to the interface as "UNIX" systems. IBM's OS/390 is an example of an operating system that includes a branded UNIX interface."
So, you are showing your ignorance by grousing over nil null's comment.
I am quite aware that everyone is not a nerd/geek.
However, slashdot is 'news for nerds' - which implies that discussions here should follow that trend.
My concern is I am seeing quite a few people posting about computing who have no business making their opinions known because they speak from a standpoint of ignorance. Additionally, the vast majority of these folks seem to post anonymously.
Unless they have used unix/linux *and* windows for an extended period of time (I loaded Slackware back in 1992, and have been a dos/windows weenie since the 80s) how can they comment on it and have any credibility?
The next question that follows is, what is their agenda?
If you are a nerd, what the heck are you doing running windows anyway?
You should have several linux boxes and do all of your configurations via VI(M) or EMACs on your primary box, and use perl/expect.pm and/or Cfengine for remote administration of the others.
Why is everyone so enamored with GUIs anyway? It takes me almost no time at all to configure an application or service using a text editor, whereas I spend hours searching through dross in GUIs to get one simple thing done in windoze.
The key difference is you have to know something about the underlying operating system - which I thought defined 'geekitude' and general 'nerdishness' best... I guess modern geeks can't handle groking details anymore. It is a sad state of affairs - a sad time we live in...
Actually, that is a misnomer. Linux is not just a monolithic kernel, it is a hybrid kernel that while containing monolithic structures, also allows the dynamic loading of drivers as needed. You also mention how much 'time' is spent configuring a linux box versus tuning an NT or XP machine to get the same functionality (doubtful - but will give you that for the sake of argument) - it takes me about 15 minutes to run through the linux kernel script and start the compile...after that my time is my own while the machine does all the work. After I have a new kernel it takes about 2 minutes to put it in place and set up the hooks in Lilo to load it. On the other hand, taking a microsoft kernel and configuring it can take several days...and god help you if you corrupt the registry in the process.
The real test is how stable a system is. I have linux machines - that have had the kernel rebuilt btw - that have been running constantly for 6 months or more. I have rarely seen a Windows box with an uptime of greater than a few weeks, and those that are simply don't do any real work.
Throw on top of that all of the cost associated with maintaining Windows liscenses, and all of the spyware that is built into XP and.NET - a particular problem I would think for government and businesses that deal with secret/proprietary information - and you have a perponderance of issues that I frankly, don't want to deal with in my operating system.
If you are fine with all of those limitations - that is alright. However, don't try to feed me dog food and call it filet mignon. Each operating system has its uses. For technically savvy users - like most of the slashdot community - the selection criteria is clear, in mission critical environments I would take Linux over any Microsoft product hands down.
On a P100 (back in the day) it only took 4 hours to recompile the kernel.
Now, I recompile on a faster machine with all of the settings tweaked for the machine I am compiling for. Then I move the binary over.
As for speed - most kernels that ship (except for Mandrake - that may have changed ) are set up to only use the least common denominator 80386 chipset commands. One of the best things you can do for a P100 is to make the kernel take advantage of the 80586 (pentium) command set. Additionally, its not all about speed as much as it is about saving space in ram - particularly on an old machine with little memory. If you are using one of these old machines as part of a cluster, or as a group of servers that have specific jobs, then the small speed and memory enhancements will add up over time.
Finally, I have more than one machine, so I can dedicate one machine to recompilation, while I continue work on another - its a little known concept called multi-tasking, and people can do it too.
1. When I throw together an old box (I thought I mentioned most of what I have are old p100s...maybe I didn't) I tune the kernel specifically for it and then leave it alone. It probably runs this way for 3 or more years (who 'upgrades' a p100 for cryin' out loud...?). When I build a machine for a specific purpose I get what I need and put it together - and that is how it stays for long periods of time. If you are upgrading your hardware 'frequently', then you must have more money than sense. Recompilation is a good thing, you are just used to be getting fed the 'dog food' so much you think its filet mignon.
2. Your second statement is pure FUD. Having used both windows and linux for many years now, there is no contest: Linux documentation is infinitely more useful than anything I have seen from Microsoft. With Linux documentation I get everything I need to install, configure, and maintain a system - as well as a wealth of information on the web via the Linux Documentation Project or directly from the application developers in most cases. All the basic unix commands are documented via man pages (simply type 'man ' in a command line shell, and you will get a concise listing that helps you specify the parameter list, as well as explains how the application works). Linux documentation runs the gamut, I will admit that; but, by and large, it is more useful. Microsoft documentation is written for an 8th grader, and is not useful for anything more difficult than changing your screen colors; I don't particularly care for Microsoft's paternal "We know what is best for you" attitude which shows everytime I try to find anything useful from a technical standpoint.
Slashdot Mirror
Cduffy,
I would love to take your friend up on it, since we have had some difficulty finding system admins. However, my company is not authorizing outside hires atm.
Sorry to get your hopes up.
Isn't this essentially what ethernet does? Two channels broadcast on at the same time, and nothing responds, each one sets a pseudo random time to rebroadcast...the internet runs on this system and doesn't die - why can't radio be set up along similar lines? (heck we used to use repeaters for our radios back in my army days).
Additionally, we are doing wave division multiplexing with T1 circuits, as well as fiber optics using multi spectrum...not to mention DSL across a 2 wire circuit layered on top of voice...
Given all of that, I think it is feasible to build technology to make the concepts discussed in the article reality.
I work in the 'real world' and here is the experience we are looking for:
.NET
Linux/Unix - both system programming and system administration experience. Show us how you would automate various features, and integrate different systems together to get real work done quickly.
VOIP and Telephony - convergence is not just a buzz word.
Java/CGI/XML - web enabled application development is a must. No one I am talking to is considering
A plus is experience using Perl/Tk, TCL/Tk, C++ (gcc), emacs, vi, awk, sed, and shell scripting.
Things that will not get you hired:
Primary Microsoft experience; Microsoft certifications mean nothing in our space. I've lost count of how many microsofties come in looking for work, and are totally lost in the datacentre.
MBA - you would be surprised at how many folks think 'system administrator' means 'managing people'; if you don't have a technical background, forget it.
HERE is what a guy who worked for microsoft has to say about the passport login system hotmail and other MS and partner sites use.
Probably why most people balk at having to create accounts and login to these sites.
I think some folks have a misapprehension that a CS degree only equates to being a programer.
Disabuse yourself of that fallacy, and you will be set.
Case in point: Me.
I started out of college NOT AS A PROGRAMMER - but as a system administrator (I parlayed my brief experience with Unix in college into the job). Once in the position, I automated all of the tedious stuff, then set my sights on some of the other things around me. For example, the technical support folks were using paper tickets; a little database magic and some cgi scripts, and 'Voila!' - an automated ticketing system.
After that, I extended my knowledge - I learned other operating systems the company was using, and I also learned new programming languages. I also started developing my philosophy regarding programming paradigms for large projects by reading everything I could get my hands on and watching what went on with internal projects. Most importantly, I learned how to make disparate systems work together (systems integration).
Finally I got into a major project as a programer and applied and refined the lessons I had learned. Having administered operating systems and databases, I had a unique perspective many of my peers did not have. I volunteered for the difficult tasks, because I knew I would learn more by stretching myself than by sitting back and just coding by rote.
Pretty soon, I was project lead, and then a full blown senior developer. At this stage of the game I deal with technology and implementation issues at a high level; I write specifications, and either implement it (if its trivial) or oversee a vendor or internal team perform the implementation.
That is how a CS major can take you where you want to go; don't limit yourself by setting your expectations too high, or conversely, setting them too low. The biggest key is to just make yourself as valuable as possible to your company and you will get where you are going.
This is probably just a means for the government to collect a list of 'dissidents'.
According to government statistics, there is a direct correlation between gun ownership, people who believe in 'real' freedom (not the PC fakey kind), and anti-telemarking activists.
You have been warned...
[ The Management ]
Can anyone comment on if and how the SCO lawsuit may or may not effect linux kernel distros now or in the future?
I wouldn't throw away the earlier kernels just yet. It may not effect your average mom and pop operation, but a legal injunction would curtail alot of corporate projects that are currently using linux.
It would be nice to see someone respond who is very familiar with the kernel development and these issues.
That is a false stereotype. I have many qualities that would be considered 'computer geek' - whatever that means - and can communicate quite well, thank you.
I prefer to be considered a unique human being, as opposed to being labeled and filed into a neat little cubbyhole. Hopefully your sense of self worth makes you feel the same way. However, I could be wrong about that; you may be perfectly happy assuming the one dimensional attributes of the labels you embrace.
Enough chit-chat. Back to the grindstone...
What scares me more than the government misusing its power, is the people here who say "its too difficult - I would give my key away if a gun was to my head..."
If you are an American, you should be ashamed of yourself. This is exactly the situation we faced 200+ years ago. What if the people who founded this nation decided the sacrifice was too hard? Would they have turned turtle and spilled the beans, giving their friends and compatriots a death sentence?
While most moments of your life are spent on less than stellar events, when the time comes to be measured I only hope you will rise to the occaision, instead of slinking away like a dog.
What they need is some software that can take care of this for them as they are going through the stories; just rewrite SLASH to take care of it (I would, but I am up to my eyeballs in greased potbelly pigs atm).
Lets see if we can find the troll. Is he:
a) Microsoftie?
b) Government Employee?
c) RIAA?
d) CowboyNeal?
Why so touchy on the subject of 'Microsoft Bashing'? The ability to thumb our noses at monopolies and bad government is an american pass-time (there is plenty of prior art if you look in the history books).
Could it be you own large amounts of Microsoft stock? Do you work at the company? Or, are you just ignorant of the findings in the antitrust case?
Feel free to bash the open source community, Apple Computing, Sun, HP, IBM, or anyone else you find loathsome. Don't tell me how to regulate my expression (only CowboyNeal and the gang can do that here).
(this almost begs a new discussion on how the supreme court views freedom of expression online - you probably wouldn't like what they have to say about it)
You're not paranoid if they are really trying to get you...
"...who could get past the wall of flaming cow dung ..."
Now that describes the situation precisely.
We are having trouble now getting a system administrator for a project because HR classifies 'system administrator' as a low level 'mainframe operator' kind of job, and thus does not allow a proper paygrade for someone who is going to be managing multiple OSes and large numbers of machines with a high need to understand and manage system admin automation.
We need a superstar sysadmin; we will probably get some kid out of college who can't find his butt with both hands.
Do you study the makeup and practices of the development team as part of your analysis? Would you find it useful to know if a team favored one lifecycle methodology over another - and are there any correlations you have seen along these lines?
Bungi has a point.
/. nick. If a 'mixed' approach is needed, then the editor should write the piece, giving credit for all of the 'pieces' where credit is due.
Stories should not be altered and then attributed to a particular
I think his point is there is a perception that because our fingers aren't constantly tapping on the keyboard, we are not doing any work.
The reality is Unix/Linux system administrators are not paid what they are really worth to the organization because the suits don't understand all of the things we do on a constant basis to:
1. Develop software tools to automate various activities and save man hours for the business. Good design takes time and thought, things not available in the factory line and cube farm mentality of most organizations.
2. Mitigate disaster (natural, software [stupid programmers/viruses], or otherwise) by designing systems and networks that work - and if they fail recover automagically - or with minimal time and effort.
I consistently have the boss come by and say, "I don't see you working", or, "what are you no-goods up to?". There is no conception of the genius that is put into the systems, and its value to the company. System administrators are not interchangeable; good ones are few and far between.
Elbereth - your tirade does not become you. If you don't grok unix/linux/bsd, or don't care for our inside jokes, kindly keep it to yourself.
. html: "In 1968, AT&T's Bell Labs began development on the Unix operating system. Shortly thereafter, many hardware vendors had developed their own flavor of Unix for their particular machines, which made porting software between platforms difficult. Recognizing the need for a standard to make porting applications between operating systems simpler, the IEEE developed POSIX in 1988 and in 1990 it became International Standard ISO/IEC 9945-1:1990 [14]. The United States Government has adopted POSIX as a Federal Information Processing Standard (FIPS 151) inspiring system vendors to do the same. The list of vendors that have announced support for POSIX includes AT&T, Control Data Corporation, Digital Equipment Corporation, Grumman Data Systems, Hewlett-Packard, Intel, IBM, Microsoft, and Sun Microsystems.
d 10_gci214309,00.html more details: "Informally, each standard in the POSIX set is defined by a decimal following the POSIX. Thus, POSIX.1 is the standard for an application program interface in the C language. POSIX.2 is the standard shell and utility interface (that is to say, the user's command interface with the operating system). These are the main two interfaces, but additional interfaces, such as POSIX.4 for thread management, have been developed or are being developed. The POSIX interfaces were developed under the auspices of the Institute of Electrical and Electronics Engineers (IEEE).
Xerithane - Linux is Posix compliant, just as BSD and the original AT&T System V. Per http://www.cs.buffalo.edu/~smgallo/SEpaper/node14
POSIX defines a standard way for an application program to obtain basic services from the operating system. It describes a set of functions derived from AT&T UNIX (System V) and Berkeley Standard Distribution UNIX (BSD). All POSIX-conforming systems must implement these functions, and programs that follow the POSIX standard use only these functions to obtain services from the operating system and the underlying hardware [14].
Although POSIX is based on AT&T and BSD UNIX, it is not an operating system itself. It simply defines an interface between software and their libraries. This allows non-UNIX operating systems to conform to the POSIX standard."
And from http://search390.techtarget.com/sDefinition/0,,si
POSIX.1 and POSIX.2 interfaces are included in a somewhat larger interface known as the X/Open Programming Guide 4.2 (also known as the "Single UNIX Specification" and "UNIX 95"). The Open Group, an industry standards group, owns the UNIX trademark and can thus "brand" operating systems that conform to the interface as "UNIX" systems. IBM's OS/390 is an example of an operating system that includes a branded UNIX interface."
So, you are showing your ignorance by grousing over nil null's comment.
I am quite aware that everyone is not a nerd/geek.
However, slashdot is 'news for nerds' - which implies that discussions here should follow that trend.
My concern is I am seeing quite a few people posting about computing who have no business making their opinions known because they speak from a standpoint of ignorance. Additionally, the vast majority of these folks seem to post anonymously.
Unless they have used unix/linux *and* windows for an extended period of time (I loaded Slackware back in 1992, and have been a dos/windows weenie since the 80s) how can they comment on it and have any credibility?
The next question that follows is, what is their agenda?
Don't worry. Palladium will come along and solve all of your 'fair use' problems.
If you are a nerd, what the heck are you doing running windows anyway?
You should have several linux boxes and do all of your configurations via VI(M) or EMACs on your primary box, and use perl/expect.pm and/or Cfengine for remote administration of the others.
Why is everyone so enamored with GUIs anyway? It takes me almost no time at all to configure an application or service using a text editor, whereas I spend hours searching through dross in GUIs to get one simple thing done in windoze.
The key difference is you have to know something about the underlying operating system - which I thought defined 'geekitude' and general 'nerdishness' best... I guess modern geeks can't handle groking details anymore. It is a sad state of affairs - a sad time we live in...
Actually, that is a misnomer. Linux is not just a monolithic kernel, it is a hybrid kernel that while containing monolithic structures, also allows the dynamic loading of drivers as needed. You also mention how much 'time' is spent configuring a linux box versus tuning an NT or XP machine to get the same functionality (doubtful - but will give you that for the sake of argument) - it takes me about 15 minutes to run through the linux kernel script and start the compile...after that my time is my own while the machine does all the work. After I have a new kernel it takes about 2 minutes to put it in place and set up the hooks in Lilo to load it. On the other hand, taking a microsoft kernel and configuring it can take several days...and god help you if you corrupt the registry in the process.
.NET - a particular problem I would think for government and businesses that deal with secret/proprietary information - and you have a perponderance of issues that I frankly, don't want to deal with in my operating system.
The real test is how stable a system is. I have linux machines - that have had the kernel rebuilt btw - that have been running constantly for 6 months or more. I have rarely seen a Windows box with an uptime of greater than a few weeks, and those that are simply don't do any real work.
Throw on top of that all of the cost associated with maintaining Windows liscenses, and all of the spyware that is built into XP and
If you are fine with all of those limitations - that is alright. However, don't try to feed me dog food and call it filet mignon. Each operating system has its uses. For technically savvy users - like most of the slashdot community - the selection criteria is clear, in mission critical environments I would take Linux over any Microsoft product hands down.
On a P100 (back in the day) it only took 4 hours to recompile the kernel.
Now, I recompile on a faster machine with all of the settings tweaked for the machine I am compiling for. Then I move the binary over.
As for speed - most kernels that ship (except for Mandrake - that may have changed ) are set up to only use the least common denominator 80386 chipset commands. One of the best things you can do for a P100 is to make the kernel take advantage of the 80586 (pentium) command set. Additionally, its not all about speed as much as it is about saving space in ram - particularly on an old machine with little memory. If you are using one of these old machines as part of a cluster, or as a group of servers that have specific jobs, then the small speed and memory enhancements will add up over time.
Finally, I have more than one machine, so I can dedicate one machine to recompilation, while I continue work on another - its a little known concept called multi-tasking, and people can do it too.
1. When I throw together an old box (I thought I mentioned most of what I have are old p100s...maybe I didn't) I tune the kernel specifically for it and then leave it alone. It probably runs this way for 3 or more years (who 'upgrades' a p100 for cryin' out loud...?). When I build a machine for a specific purpose I get what I need and put it together - and that is how it stays for long periods of time. If you are upgrading your hardware 'frequently', then you must have more money than sense. Recompilation is a good thing, you are just used to be getting fed the 'dog food' so much you think its filet mignon.
2. Your second statement is pure FUD. Having used both windows and linux for many years now, there is no contest: Linux documentation is infinitely more useful than anything I have seen from Microsoft. With Linux documentation I get everything I need to install, configure, and maintain a system - as well as a wealth of information on the web via the Linux Documentation Project or directly from the application developers in most cases. All the basic unix commands are documented via man pages (simply type 'man ' in a command line shell, and you will get a concise listing that helps you specify the parameter list, as well as explains how the application works). Linux documentation runs the gamut, I will admit that; but, by and large, it is more useful. Microsoft documentation is written for an 8th grader, and is not useful for anything more difficult than changing your screen colors; I don't particularly care for Microsoft's paternal "We know what is best for you" attitude which shows everytime I try to find anything useful from a technical standpoint.
Yeah, and the counter on that page:
00000002
Probably Steve and Bill...