On Fri, 25 Jun 2004 14:54:19 -0500, Bill Denton wrote:


I have seen quite a few Linux admins who had to use cheat sheets for even
the most rudimentary Win server tasks. And during the dot.com bust I saw
more than one Linux admin deliberately allow their Windows boxes to
deteriorate so they couild make the Linux boxes look better, and thus
preserving their jobs.

During the .com boom, there where plenty of completely unqualified people
working in IT. People often needed warm bodies to support their 1, 2,
4 business plans. The scales heavily lean toward WIN admins, IMO, but I
have no doubt that there have and are unqualified Linux/Unix admins out
there too. In fact, I've met unqualified HP/UX admins too. A rotten egg doesn't
spoil the whole batch unless you attempt to cook with it.


I have seen applications crash NT workstation and server four or five times,
and I've crashed Win2K Professional twice; once with Flight Simulator. A
couple of years ago I worked for a software company, and a test suite run of
one of our applications brought down two Linux boxes and one Unix box.
Everything can crash.

Granted, but if an application crashes, it's an OS bug. There are many
ways to "bring down" a system, but not all of these are bugs. In many
cases, it's improper configuration or hardware for a task.


And I saw a Win2K server mysteriously begin going BSOD, for no observable
reason. It looked just like a software crash. I worked in the IBM building
in downtown Chicago, where you would expect the power to be good, but it
turned out we were getting power sags which were crashing the machine.

This is exceedingly common. The more machines you get on a circuit, the
dirtier the power is going to become. People often rush to purchase surge
protecters but fail to realize the brownouts are actually far, far worse
for their computer. Not to mention, very common. They often go
completely unnoticed but cause bit-flip errors or even physical gate
damage to some components. I could go on, you I think you get the point.

It
was plugged into the same outlet with a workstation which never had a
problem.

Different CPUs, chipsets, power supplies, and yes, even computing trends,
can all make a significant difference. Not all computers are created
equal.

I did some testing, and discovered that the sags were long enough
to drop the server, but not long enough to effect the workstation. I put in
a UPS; no problem. But as I said, it looked just like a software problem.
How many other hardware problems get blamed on the OS?

I'm sure many.


It's not a matter of which will do the job, it's a matter of which will do
the job best. And there are things that Win will do better than Linux and
vice-versa. And better is not just a matter of benchmarking: in some small
towns you might find 10 Win administrators and zero Linux administrators. In
that case, Linux is totally worthless. It's all a matter of matching the OS
to the need.

Well, my statement assumed everything being equal. But you are right.
I'm a strong believer in the best tool for the job. If your room of a
thousand monkeys only know typewriters, you don't give them word
processors and hope for the best. You have a valid point here.


And I hate to tell you this, there were LAN's long before the Internet
became "prime time". I did my first Windows For Workgroups (NETBUI) network

Done that too. Ohh...you remember arcnet? Hehe. Those were the days.

in 1992, the Internet did not begin to achieve any sort of mass penetration
until 1996 or so. While the majority of the servers may run Unix/Linux, most
of the outbound data quickly goes through a router onto a Windows network,

Well, before the big days of the net, most were military, dod, and
university users, so it was still mostly unix. Granted, as it grew,
windows networks were certainly plugged in.


I agree with you about the increased vulnerability of Windows, but a case
could also be made that the associated protocols, which were designed by
Unix guys, were poorly engineered. Had they been better designed the spam
problems would not exist. I don't make that argument, but if you talk
about Windows vulnerabilities, you also have to consider lacadasical
engineering.

Hehe. I don't think that's fair. IPv4 works rather well for what it was
designed to do. It's just the IPv4 has grown well beyond it's original
design. This is why IPv6 exists. This is why everyone is wanting the
Interent to transition to IPv6. It addresses most every complaint of
IPv4. So, I don't think it's fair to blame the designers if the market
refuses to adopt the correct technology.

Having said that, spam is actualy an issue of the SMTP protocol and not
the IP protocol. Granted, better facilities in IP may of helped. But
let's face it, the current SMTP protocol (and associated RFCs) would be
hard pressed to become more spam friendly. If you insist on pointing a
finger, feel free to point a finger at the right group.

Cheers!

Greg