Volume 2, Number 5 -- May, 1997
| |
Volume 2, Number 5 -- May, 1997 |
 
|
Pelican Smith
Someone asked Bruce Lee once what his most advanced technique was. He
replied that the most advanced technique is a well executed basic
technique. Beginners may often mistake complexity for effectiveness,
but veterans know better.
There is a stigma attached to both Unix and Microsoft. Unix was developed
by geniuses for geniuses in the infamous Bell Labs. It is the stuff of CS
legend. Whoever used Unix in the early days was most likely a physics
major or engineer - the people who would go on to define computers and
internetworking over the next twenty years.
MS-DOS came from a clever marketing scheme, not clever hacking. It was
simply revamped and repackaged, and grew to a billion dollar empire. It
was written by geniuses for idiots, people who had never seen a computer
before and needed all the help they could get.
The first Unix system I ever worked on was an ATT System V. I configured
the terminal, wrote shell scripts, and marveled at the power of
multiprocessing. I encountered many other flavors of Unix over the years,
Solaris, HP-UX, Ultrix, Xenix, and then glorious Linux, and began imagining
myself somehow better than my MS-DOS associates. I was, after all, a C
programmer in the Unix environment. I knew power.
One day my work purchased a copy of Visual Basic 3.0. By this time I was
a C++ programmer, recompiling kernels, writing Perl scripts, and creating
my own editor. I laughed at the pathetic Windoze users and their pretend
programming languages, and returned to my pipes and forks and execs.
Three years later two remarkable events occurred within the same day. I
had to move some simple shell scripts from a Dec 5000 Ultrix system over to
a Sun, and encountered a slew of errors. It seemed that the commands on
the Sun did not operate exactly as they had on the Dec. I thought this was
a crime. If "ps -ax" is the command to list all processes in Unix, then it
should be the same on all systems that claimed to be Unix. I saw it as
being equal to DIR/W working on one DOS system, but not on the next.
Attempts to standardize the various flavors of Unix have been reported for
twenty years, with no real progress.
The other remarkable event was when I sat down and started tinkering with
the Visual Basic compiler. I'd just spent two painful days writing a
simple "Conditions of Use" window for the Dec's. The user had to click
"Agree" to continue logging in, or "Disagree", which caused the sign-in to
terminate. Of couse, the same X11 program would not run on the Sun's. I
now had to write the same program for the Windows 3.1 computers. I'd never
ran Visual Basic before, but still completed the program in less than an
hour.
This was about the time when I forsaw the death of Unix.
A few years later my company discovered the Internet, and my job changed
from programmer to network consultant. I built web servers and mail hosts
and DNS's and news servers and all the rest. The process was always the
same: Spend a couple hours tinkering with config files and Makefiles,
compile the program, install the program, and then do it again in a couple
mounths after the system crackers learned how to hack apart the programs.
I was spending countless hours explaining simple Unix administration to
people at other sites, begging them to check their system logs, teaching
them to use vi, and all the time realizing that they simply weren't going
to do it.
When it was decided to migrate Internet services to Windows NT, I decided
to jump on board. It was the best decision of my career.
Windows NT 4.0 now packs all the power of Unix, and is infinately more
convenient to use. Most people, already familiar with the Windows
interface, are not afraid of an NT server. It's all the same: Drag and
Drop, Cut and Paste, Double click, hey, a kid could do this.
When it comes to security, NT has taken it's lumps, but I still argue that
it can offer more Internet services more securely than any Unix system. If
nothing else, the source code for NT is not freely available, which makes
it more difficult to hunt for bugs in the code. There's no real need for
Unix's biggest security hole - the telnet daemon (although one is available
for NT).
Better work is getting done faster. People are happier. Best yet, I now
have eager trainee's, as apposed to the yawns and confused looks I used to
get when I tried to explain the sendmail.cf file.
Unix became popular because it was powerful and easy to use. It has not
kept up with modern technology. This is the death of Unix.
This month's top 5 web picks: |