[comp.os.coherent] What's the Catch?

rose@galtee.cs.wisc.edu (Alice Waters) (04/06/91)

Before folks get too excited about the prospect of porting big things like
TCP/IP or X11R-anything to Coherent, allow me to point out one little thing
that might rain on that parade a wee bit-- the compiler only supports small
model.  That means 64K code and 64K data, max.  Now, you could do some creative
things with a ram disk to make things work, but it isn't going to be a matter
of just changing a few defines in the code to get big things to compile-- its
going to be the sort of port where you really get your hands dirty.  I 
personally would want to have a good assurance that a 32-bit version of the
operating system was not going to come out the day after I finished a port like
that, but that might just indicate a character flaw on my part.

-- 
	Scott Rose
	rose@cs.wisc.edu
	(608) 238-3801

winans@sirius.mcs.anl.gov (John Winans) (04/08/91)

In article <1991Apr5.163104.28033@spool.cs.wisc.edu> rose@galtee.cs.wisc.edu (Alice Waters) writes:
>Before folks get too excited about the prospect of porting big things like
>TCP/IP or X11R-anything to Coherent, allow me to point out one little thing
>that might rain on that parade a wee bit-- the compiler only supports small
>model.  That means 64K code and 64K data, max.  Now, you could do some creative
>things with a ram disk to make things work, but it isn't going to be a matter
>of just changing a few defines in the code to get big things to compile-- its
>going to be the sort of port where you really get your hands dirty.  I 
>personally would want to have a good assurance that a 32-bit version of the
>operating system was not going to come out the day after I finished a port like
>that, but that might just indicate a character flaw on my part.

Perhaps it is my age showing.  But I remember when 64K for both code AND data
was "more than anyone would ever want."

*SIGH*     Ahh.  The good ol' days  :'(


But times do change & as an observation (that I am sure MWC has made too) a 32 
bit version of Coherent for less than $500 would probably become THE OS of the 
90's.  Even if there is no source available for the thing.  Just don't loose
the loadable device driver stuff and I think us hacker types will stay happy.

--
! John Winans                     Advanced Computing Research Facility  !
! winans@mcs.anl.gov              Argonne National Laboratory, Illinois !
!                                                                       !
!"The large print giveth, and the small print taketh away"-- Tom Waits  !

cy5@cunixa.cc.columbia.edu (Conway Yee) (04/10/91)

In article <1991Apr8.155202.27275@mcs.anl.gov> winans@sirius.mcs.anl.gov (John Winans) writes:
>But times do change & as an observation (that I am sure MWC has 
>made too) a 32 bit version of Coherent for less than $500 would
>probably become THE OS of the 90's.  Even if there is no source 
>available for the thing.  Just don't loose the loadable device 
<driver stuff and I think us hacker types will stay happy.

Why?  For most UNIX sites, the suits demand AT&T UNIX.  Commercial
software is not sold for Coherent or even most flavors of UNIX.  
Instead, binaries only & vendor specific (i.e. SCO) packages are the
rule.  A 32 bit Coherent would not change this.  Most if not all 
commercial software vendors will not come out with Coherent specific
versions.  Besides, a 32 bit version of MINIX exists and you get the 
source.  Few would argue that the existance of a 32 bit MINIX has 
caused it to become a runaway bestseller.  In order to call a 32 bit
Coherent THE OS of the 90's, you have to justify why it would be better
than:

1) the many flavors of AT&T UNIX (you can claim cost advantages 
   but not third party support)
2) MINIX which comes with source and is similarly priced.

					Conway Yee, N2JWQ
yee@ming.mipg.upenn.edu    (preferred)             231 S. Melville St.
cy5@cunixa.cc.columbia.edu (forwarded to above)    Philadelphia, Pa 19139
yee@bnlx26.nsls.bnl.gov    (rarely checked)        (215) 386-1312