Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!usc!apple!snorkelwacker!mit-eddie!rutgers!columbia!cunixf.cc.columbia.edu!cunixb.cc.columbia.edu!es1
From: es1@cunixb.cc.columbia.edu (Ethan Solomita)
Newsgroups: comp.sys.amiga
Subject: Re: Stability of Commodore/Amiga
Keywords: Unix, Amiga, OS/2, DOS
Message-ID: <1990Mar5.215815.159@cunixf.cc.columbia.edu>
Date: 5 Mar 90 21:58:15 GMT
References: <476087196@deimos.cis.ksu.edu> <19000039@attcc.UUCP> <2840@mtuni.ATT.COM> <676@xdos.UUCP> <3881@nmtsun.nmt.edu> <48ffd21f.db93@edsel.engin.umich.edu> <3883@nmtsun.nmt.edu> <3093@bingvaxu.cc.binghamton.edu>
Sender: usenet@cunixf.cc.columbia.edu (The Network News)
Reply-To: es1@cunixb.cc.columbia.edu (Ethan Solomita)
Organization: Columbia University
Lines: 30

In article <3093@bingvaxu.cc.binghamton.edu> consp11@bingvaxu.cc.binghamton.edu.cc.binghamton.edu (Brett L. Kessler) writes:
>In article <3883@nmtsun.nmt.edu> dksnsr@nmtsun.nmt.edu (Dr. Mosh) writes:
>IMHO, contiguous memory is not the way to go.  If I have a 300k
>program on a contiguos-memory-based machine (ie: Mac) that has a 290k
>area free and a 50k area free, the program still cannot be loaded,
>becuase the areas are not contiguous.
>
>Linked lists (and related structures) are a much more efficient way of
>allocating memory.  True, it incurs a bit of overhead and requires
>some work by he programmer (of the OS, that is), but the 300k program
>example from before would load just fine into an Amiga - it doesn't
>require contiguous blocks to know what it's doing.
>
>+------///-+------------------| BRETT KESSLER |------------------+-\\\------+
>|     ///  |         consp11@bingvaxu.cc.binghamton.edu          |  \\\     |
>| \\\///   |              consp11@bingvaxa.BITNET                |   \\\/// |
>|  \XX/    |              (PeopleLink)  B.KESSLER                |    \XX/  |
>+----------+-----------------------------------------------------+----------+


	As you say, the program has to be designed in such a way as to
load in pieces. If it attempts to load as one large 300K alloc, it
will fail. What I was wondering was is there a way to transparently
have the OS deal with that. Perhaps it would require an MMU, but is it
possible?
	-- Ethan

Ethan Solomita: es1@cunixb.cc.columbia.edu
Compu$erve    : 70137,3271
Anyone giving away Amigas or Sharp Scanners???