comp.lang.ada
 help / color / mirror / Atom feed
From: "Marin David Condic" <mcondic.auntie.spam@acm.org>
Subject: Re: Modern copy of GNAT for OpenVMS?
Date: Sat, 7 Dec 2002 10:26:31 -0500
Date: 2002-12-07T15:31:26+00:00	[thread overview]
Message-ID: <ast48e$4v2$1@slb5.atl.mindspring.net> (raw)
In-Reply-To: 9yPmFNnjB6V5@eisner.encompasserve.org

Very true. It is a problem that OS's all have different ways of doing
things. In some ways, its "Innovation" and in other ways its a handicap.

Back in the olden days, you could count on your garden variety user-style OS
(as opposed to something embedded or otherwise specialized) providing you
with basically sequential, random and text IO with some version of a
teletype as your user interaction device. At that time, what Ada provided
was sufficient to accomplish portability across a multitude of platforms -
possibly excepting IBM mainframes where the terminals behaved dramatically
different and character sets were not compatible with the standard. Now
you've got a variety of windowing systems and networking systems and file
systems that may each have their own advantages, but it makes any sort of
portability difficult. You either fully utilize what is available to you or
you have to opt for the least common denominator - which may put you back at
a glass teletype and fairly primitive files - with maybe networking since
Sockets are fairly common - yet often with subtle differences.

So given that portability of user-style apps is pretty much impossible
without making the app very primitive, should we be so concerned about it?
Standardization is nice, but unachievable on some levels, so maybe it ought
not to be such an overwhelming concern for Ada. At least not where it might
involve big libraries of OS related things? (IOW, try to arrive at a common
subset of features, but allow divergence for different implementations.
Example: Make a root package for some subset of OS features and allow
implementations to make their own child packages for extended features with
the caveat being that if you use them, you are not portable.)

MDC
--
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jast.mil/

Send Replies To: m c o n d i c @ a c m . o r g

    "I'd trade it all for just a little more"
        --  Charles Montgomery Burns, [4F10]
======================================================================

Larry Kilgallen <Kilgallen@SpamCop.net> wrote in message
news:9yPmFNnjB6V5@eisner.encompasserve.org...
>
> The problem is that what is "normal" for one operating system is not
> at all "normal" for another.  Even the alleged DLM port to Tru64 is
> quite different than the original on VMS.





  reply	other threads:[~2002-12-07 15:26 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-12-04  0:35 Modern copy of GNAT for OpenVMS? Zane H. Healy
2002-12-05 14:00 ` Vincent DIEMUNSCH
2002-12-05 14:30   ` Stephen Leake
2002-12-05 14:56     ` Larry Kilgallen
2002-12-05 17:23   ` Jeffrey Carter
2002-12-05 17:12     ` Larry Kilgallen
2002-12-06 12:48     ` Marin David Condic
2002-12-06 12:56       ` Larry Kilgallen
2002-12-07 15:26         ` Marin David Condic [this message]
2002-12-06 20:37       ` Frank J. Lhota
2002-12-06 21:28         ` Larry Kilgallen
2002-12-07 15:33           ` Marin David Condic
  -- strict thread matches above, loose matches on Subject: below --
2002-12-05 17:45 Caldwell Ian
2002-12-06 20:42 ` Tapani Rundgren
2002-12-06 20:57   ` Stephen Leake
2002-12-07 15:29     ` Tapani Rundgren
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox