[net.micro] Advice on micros needed

paul@oddjob.UUCP (Paul Schinder) (04/17/85)

A friend and I are thinking of buying micro computers for home use, and were
wondering if anyone could point us to machines that would meet our needs.
Ideally, the micro would have 1. good floating point capabilities (for
scientific program development and possibly some actual running; a Fortran
77 compiler is essential), 2. good graphics capabilities (ideally, Tektronix
4014 emulation) with software to make hardcopies on an attached printer as
well, 3. good communications software, so that transferring files between
the available mainframes and the micro is easy (it would be best if the
program(s) needed on the mainframe was widely available or easy to write,
since the mainframes in question could range from VAX to CRAY X-MP), 4.  a
good local editor, the more emacs-like, the better.  A dream machine would
be an IBM PC/AT with 20Mb hard disk, math coprocessor, and Xenix, but that
is out of our price range of ~ $3K (people associated with the University
get a substantial discount on a wide range of micros, so this limit is
equivalent to about $5K retail).

It has boiled down to three general possibilities:

1.  Forget the floating point and local program development capabilities and
buy a cheap (Commodore class) computer which does the communications and
graphics; it would be used mainly as a smart terminal with local hardcopy
capabilities.  Wait a few years to get a "real computer".

2. Get a 512K Macintosh.  I understand that Fortran compilers are now
available, but haven't heard much about how good they are (are they Fortran
77 standard?).  The great advantages of the Macintosh are that there is a
large amount of free software offered by the comp center here, that it is
reasonably portable, and that there are a great many 128K and 512K Macs
around campus, so advice is readily available.  MacTerminal or kermit and
VersaTerm take care of the communications and Tektronix emulation.  From
what I've read on the net, however, it seems that it's hard to actually
program the thing, and the floating point capabilities may be too weak.  We
fear that this option may actually turn out to be a very expensive smart
terminal, rather than a "real computer", and so are hesitant to go out and
get one.

3.  Buy a "real computer".  Strain and spend a little bit more and get
something which can be built into the dream machine (the lowest cost IBM
PC/AT and a cheap printer, for instance).  The HP Integrated Personal
Computer is a very intriguing option (featuring our favorite operating
system), but what kind of compilers are available for it and how are the
communications and graphics?

Any comments or recommendations about micros would be greatly appreciated.
I'd rather not hear comments like "you should be programming in Pascal or C
or ...".  I'm very fluent in C and Fortran, and in my opinion Fortran is a
much better language for scientific programming.  Fortran is also available
on every scientific computer; other languages are not.  Please respond by
e-mail; if there is sufficient interest, I will post a summary to the net.

Thanks very much.

-- 


				Paul Schinder
				Astronomy and Astrophysics Center
				University of Chicago
				uucp: ..!ihnp4!oddjob!paul