home *** CD-ROM | disk | FTP | other *** search
- Path: phoenix.rhein.de!yaps!arno
- From: arno@yaps.rhein.de (Arno Eigenwillig)
- Newsgroups: comp.sys.amiga.misc
- Subject: Re: OS features
- Message-ID: <hmAVx*Y3f@yaps.rhein.de>
- Date: Tue, 09 Jan 1996 15:06:57 +0100
- References: <92747544038@PAPA.NORTH.DE> <4b3h9s$1st@alterdial.UU.NET>
- <2152.6561T63T2136@cycor.ca> <4b7i18$si1@vixen.cso.uiuc.edu>
- <oj6raxxrr0o.fsf@hpsrk.fc.hp.com> <13213431@sourcery.han.de>
- <4cpmlv$obe@ousrvr3.oulu.fi> <4cqqfq$khf@serpens.rhein.de>
- <1850.6581T1132T2235@idefix.wu-wien.ac.at> <4cs6ri$no8@serpens.rhein.de>
- Organization: Yet Another Private Site in Meckenheim, Germany
- X-Copyright: This article may not be distributed on a CD-ROM
- or in printed form without prior written consent of the author.
- X-Newsreader: Arn V 1.04
-
- Hi.
-
- I suppose nobody cares, but my opinion on memory-protection for
- AmigaOS is as follows:
-
- AmigaOS is a great OS. That is, a great implementation of an elegant
- concept based on certain, let's say, axioms. AmigaOS has evolved quite
- nicely from a bare-bones release 1.0 to a truly powerful release 3.1.
- This evolution has taken place on basis of these axioms. New unifying
- layers have been laid over existing ones, like GadTools over Intuiti-
- on.
-
- The one axiom relevant here is:
-
- All memory shall be available
- from every context at any time
- using a single address.
-
- I think I need not elaborate about the ubiquitous effects of that
- axiom. Although there are remnants of pre-1.0 attempts to overcome
- this (MEMF_PUBLIC, mn_Length, ...), over 10 years of existing prac-
- tise have set this axiom into stone.
-
- For some time now, again and again plans have been proposed to stretch
- the OS beyond that axiom. Such plans always lead to two problems:
-
- 1) How to we stay compatible?
- 2) What other OS features need to be re-designed to fit to memory
- protection.
-
- No. 1 always is intended to be solved by treating old software as a
- special case, giving it a common address space without mutual protec-
- tion.
-
- No. 2 usually is not elaborated further, but is nevertheless very
- important. Virtually all non-trivial interactions of applications and
- OS as well as communication among cooperating applications need to be
- re-designed or would infer severe performance penalties.
-
- Therefore, I believe that a hypothetical new AmigaOS with memory
- protection (and, less problematic, virtual memory) would be an ugly
- kludge, because it either cannot offer full memory protection and has
- to make compromises to keep old semantics of the programming model, or
- will have to offer most operations in several ways, depending on the
- level of memory protection applicable for the individual application.
-
- The problems of having all exiting drivers migrate to the new system
- have already been mentioned.
-
- AmigaOS is nice within its design axioms. Let's not try and stretch it
- beyond them. DOS and Windoze have already demonstrated that this leads
- to bloat and kludginess.
-
- If we need a very tight integration of memory, files and arbitrary
- high- and low-level I/O streams, let's take Un*x right away and not
- half-heartedly make AmigaOS a bad imitation.
-
- -- __
- __/// Arno Eigenwillig /\ <arno@yaps.rhein.de> \/ PGP key available.
- \XX/ V+49-2225-5870 /\ <Arnooo @ #amigager> \/ MIME 8bit welcome.
-