home *** CD-ROM | disk | FTP | other *** search
- Path: EU.net!sun4nl!xs4all!usenet
- From: jtv@xs4all.nl (Jeroen T. Vermeulen)
- Newsgroups: comp.sys.amiga.misc
- Subject: Re: OS features
- Date: Sun, 14 Jan 96 23:17:11
- Organization: Leiden University, Mathematics & Computer Science, The Netherlands
- Message-ID: <19960114.7B6E668.149C3@asd06-10.dial.xs4all.nl>
- References: <DL5t8z.Ktx@inter.NL.net>
- NNTP-Posting-Host: asd06-10.dial.xs4all.nl
- Mime-Version: 1.0
- Content-Type: text/plain; charset=iso-8859-1
- Content-Transfer-Encoding: 8bit
- X-NewsSoftware: GRn 2.1 Feb 19, 1994
-
-
- In article <DL5t8z.Ktx@inter.NL.net> hguijt@inter.NL.net (Hans Guijt) writes:
-
- > What I wanted to say is just this: don't get too obsessed by memory
- > protection. I would be very happy with a less than complete memory
- > protection scheme (just preventing writes to specified address ranges for
- > instance). Unrelated tasks could be protected from each others damaging
- > actions quite easily, and even if the scheme is incomplete it could stop 90%
- > of all problems. A 100% solution is probably not achievable and perhaps not
- > even acceptable for reasons of speed and compatibility.
-
-
- I'll second that. It's just too easy to pretend that memory protection is such
- a 100% safety feature. I've had UNIX machines crash from the login prompt, I've
- seen the GUI freeze completely so you can't even activate a window. Right now,
- mysterious configuration problems are making one in eight or so of our UNIX
- machines (all completely identical running the exact same software) completely
- useless and there's nothing you can do about it. Processes just freeze on
- startup and don't leave when you `kill' them. Before that, we had machines
- refusing to open shells for a particular user and the problem was never solved
- until it went away of its own free will.
-
- At some point the worst problems aren't the ones you can solve with better
- memory protection anymore; in reality all these protection features need to be
- a tradeoff of safety vs. probability, efficiency, benefit and code size. "Full"
- protection bloats the OS, and bloated code will contain more bugs no matter what
- you do.
-
- Memory protection will probably be the death of Windows NT eventually, as more
- and more code gets integrated into the OS with too many privileges. I hope the
- Amiga doesn't get caught in the same trap.
-
- --
- ============================================================================
- # Jeroen T. Vermeulen \"How are we doing kid?"/ Yes, we use Amigas. #
- #--- jtv@xs4all.nl ---\"Oh, same as always."/-- ... --#
- #jvermeul@wi.leidenuniv.nl \ "That bad, huh?" / Got a problem with that? #
-