Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!unmvax!pprg.unm.edu!hc!ames!amdcad!sun!pitstop!sundc!seismo!uunet!mcvax!unido!ztivax!tumuc!lan!ocker
From: ocker@lan.informatik.tu-muenchen.dbp.de (Wolfgang Ocker)
Newsgroups: comp.sys.atari.st
Subject: Re: Multitasking: What is available?
Message-ID: <542@infovax.lan.informatik.tu-muenchen.dbp.de>
Date: 2 Feb 89 09:25:17 GMT
References: <8901261512.AA11110@ucbvax.Berkeley.EDU> <1574@psu-cs.UUCP>
Reply-To: ocker@lan.informatik.tu-muenchen.dbp.de (Wolfgang Ocker)
Organization: Inst. fuer Informatik, TU Muenchen, W. Germany
Lines: 53

In article <1574@psu-cs.UUCP> kirkenda@psu-cs.UUCP (Steve Kirkendall) writes:
>OS/9
>	That didn't work out.  My impression of the 68000 version of OS/9 is
>	that it would be an even better OS for an 8-bit machine with 64K and
>	a slow clock.  On an ST, it seems rather under powered.
>
>	First off, it crashes frequently -- about every hour and a half.
OS-9/68000 V2.x is really stable, provided that the implementation is well
done. There are implementations out which use the (buggy) BIOS of TOS.
This causes real problems. 
>	Second, not all of the documented features are implemented yet --
>	some floating point functions don't work, and the hard disk driver
>	doesn't perform all of the low-level calls it should (eg it won't
>	park the hard disk).
Because I don't program sohisticated arithmetic programs I don't
know missing fp functions. The missing park iocontrol for the
harddisk is a lack of the implementation you use.
>	Third, it is SLOW.  Loading and starting a program takes a rather long
>	time because OS/9 generates a CRC check value every time it loads a
>	program.  
That's true. BUT: If a program is in memory already, it can be started
immediatly (no CRC computation). Programs are kept in memory, if possible. 
A good harddisk driver uses the multiple sector transfer,
so reading a programm from harddisk consumes only some 1/10 seconds.
> Also, all programs must be written in position-independent
>	code, with all data accessed via an base register.
True, but no real problem on the 68000. And there's no restriction
on the size of stack and static data.
>	Fourth, the commands that come with OS/9 have different names ad
>	syntax than their UNIX counterparts.  This didn't matter much 5 years
>	ago, but it matters to me now.
True, too. But there are shells available which allow aliasing. So
my "procs" is spelled "ps". 

May I summerize: The quality of OSK is highly dependant on the quality
of the implementation. Watch out for a better one! (Others are
available.) Many many UNIX tools have been ported and are free
available. 

/// Wolfgang Ocker

P.S.: Porting from UNIX to OSK is not so difficult as many people
believe. The port of Elm 2.1 beta took me about two days (the XENIX
port took the same amount of time ...)

P.P.S.: Im a really great MINIX fan! Thanks Andy!

/// Wolfgang
-- 
  |  Wolfgang Ocker          |  ocker@lan.informatik.tu-muenchen.dbp.de  |
  |  Lochhauserstr. 35a      |      pyramid!tmpmbx!recco!weo (home)      |
  |  D-8039 Puchheim         |     Technische Universitaet Muenchen      |
  |  Voice: +49 89 80 77 02  |          Huh, What? Where am I?           |