home *** CD-ROM | disk | FTP | other *** search
- Path: fc.hp.com!news
- From: koren@hpsrk.fc.hp.com (Steve Koren)
- Newsgroups: comp.sys.amiga.misc
- Subject: Re: Why I switched from the Amiga
- Date: 15 Jan 1996 09:55:44 -0700
- Organization: HP Fort Collins Site
- Sender: koren@hpsrk.fc.hp.com
- Message-ID: <oj6d98lbdnz.fsf@hpsrk.fc.hp.com>
- References: <4bngub$2k4@madrid.visi.net> <4bnkbk$1mmi@news.doit.wisc.edu>
- <820489831.13686@alladin.demon.co.uk> <DKqMnv.IHx@eskimo.com>
- <peterk.0jya@combo.ganesha.com>
- NNTP-Posting-Host: hpsrk.fc.hp.com
- In-reply-to: peterk@combo.ganesha.com's message of 11 Jan 96 01:01:23 MEZ
- X-Newsreader: Gnus v5.0.9
-
-
- peterk@combo.ganesha.com (Dr. Peter Kittel) wrote:
-
- > If you program cleanly from start and don't just "experiment with code",
- > you won't crash your Amiga so often.
-
- Yeah, sure you won't. I spent a few hours yesterday writing some
- functions to manage background tasks in a particular shared library. I
- wrote them as cleanly as I could. I probably had 2 or 3 dozen system
- crashes before I got it working, and I considered that pretty good.
-
- There are many subtle details that must be kept proper track of for some
- types of Amiga programming (we unfortunately have an API which makes
- CPU-register level information visible to the programmer), and if you
- don't, you *will* crash. Granted some types of programming you can do
- without many crashes, but certainly not all. And "Not many crashes" is
- still orders of magnitude more than one must contend with on Unix or NT.
- I sometimes have one reboot every 2-3 *minutes* when developing some
- types of low level software, and I've been writing Amiga software for
- longer than most people have even owned an Amiga.
-
- If Amiga programming is a pain in the butt because the machine crashes
- all the time, that is just one more reason why people will not do it.
- Already there is the small matter of a vanishingly small market share,
- noncompetitive CPU performance, and very low end standard graphics. We
- cannot afford to add straws to the camel's back here, I think. All
- these things can be addressed, but there is no surplus of time, and I
- sure don't want to pretend they are not problems. The absolute single
- worst thing for the Amiga would be to pretend these are not problems and
- thus fail to address them.
-
- > accesses, but there's another big issue: For memory protection, the
- > whole messaging system of the OS would have to be changed e.g. to a
- > system where messages are copied instead of just passing pointers.
-
- Yes, but there is still benefit to be obtained by leaving messages as
- they stand and protecting executables and indicated private data areas.
- It would obviously not be as effective as protecting everything, but it
- would be quite a lot better than matters as they stand now. I can
- validate incoming messages. I cannot stop another program from writing
- garbage into my executable.
-
- - steve
-