#Unix 486 woes
I called Snitily's (sp?) r.e. getting gcc and X11 binaries to run on top of
a "minimal" SCO or Interactive system. I was a little disappointed by the
response, which may just show my naivite about Unix. They told me that I
would probably have to get the developers' system for any Unix I picked,
since I would need the system include files if I were to compile gcc or
X11. I told them I really only wanted binaries, but there seems to be a
catch 22; the include files are proprietary, so I can't get a proper
compiled gcc unless I have the stuff to compile it on my own. I'm not sure
I buy this. I was also told that I would need to have TCP/IP to run X11r5
(though I sensed some uncertainty in this comment). Anyway, If I need to
get all this junk, I am back up to ~$1500. minimum, BEFORE I get the
binaries for X11 (the developers' modules are ~$600 to $1500 alone, on top
of the basic Unix system).
I find this incredibly frustrating. I feel compelled to learn and develop
in Unix, since I work in scientific research and programming, but I find it
damn hard to justify spending $2000. for a new file system. I do a lot of
cellular automata calculations (actually, lattice gas flow simulations),
and right now I am using Zortech's C++/386. I'm able to use all my memory
with 32-bit addressing, and by coding my inner loops in asm, my 486/33 is
running 4 times faster than the fastest Sparc at work. There is some
pressure for me to write this stuff in X11, so it will run on a wider
variety of machines; but right now I tell the people who want to run the
programs to spend $3000 on a 486 and run the code under DOS and DPMI/32.
Pardon the bitching. I'm just not used to having to work this hard to get
people to SELL me a product. I wonder if the failure to provide a
reasonablypriced Unix will eventually kill the 80×86 line. …Thanks,
Harlan Stockman…