NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c++-part1/2433

Internet Message Format | 1996-08-06 | 5.5 KB

Path: news1.ucsd.edu!usenet
From: [e-mail] (Darin Johnson)
Newsgroups: comp.edu,comp.lang.c,comp.lang.c++
Subject: Re: C or C++ for a 14-year old?
Date: 17 Jan 1996 11:34:03 -0800
Organization: UCSD Computer Science and Engineering Department
Sender: [e-mail]
Message-ID: <[e-mail]>
References: <4b30ld$lp2$[e-mail]>
    <4d4jeh$[e-mail]>
    <[e-mail]> <4ddsg4$[e-mail]>
    <[e-mail]> <4dgij2$[e-mail]>
NNTP-Posting-Host: tartarus.ucsd.edu
In-reply-to: [e-mail]'s message of Tue, 16 Jan 1996 16:12:50 GMT
X-Newsreader: Gnus v5.0

[e-mail] (Alberto C Moreira) writes:
> Why do you think the best games run under DOS ?

Because that's the largest market, and because of inertia.

It is not really for performance reasons; except that when your only
two choices are dos or windows, performance reasons do count.

There are fast games on the Mac.  Doom runs very quickly under Linux
(faster than DOS, since it can make use of 32-bit optimizations).
Amiga's had stupendous games it's whole life (and the DOS people
always called it a "game machine", sorta ironic now).  Doom, Descent,
and Dark Forces are spiffy under OS2.  And the very highest end VR
simulators run on SGI's.  These are all systems with an "intrusive"
OS.

The game developers may say the bare machine is needed for
performance, but this is rarely justified (except when they think the
only alternative is Windows).  There were (and still are probably)
fights in the Amiga camp between those developing bare-hardware games,
and those that developed games that worked with the OS.  Neither side
was a clear winner, high performance games were written by both sides.
Even the assembly-only crowd didn't have a leg up in performance.  The
bare-hardware camp however often had problems - the early games
refused to run on anything but a floppy.  Other games didn't work with
the later file systems.  Many games didn't work when the next model of
the Amiga came out, or with various incremental improvements.  There
was an absurdly high number of games that did not even have way to
stop the game, other than rebooting.  This was all defended under the
name of "performance".  The game *players* on the other hand, all
groused about the fact that they had to reboot in order to play a game
or to stop playing (yes, many Amiga owners used their machines to do
real work).  The games that worked with the OS worked just as fast
and most also continued to work even with new versions.

Note that System Shock had an optional 640x480 svga mode, but it runs
pretty slowly, even on relatively fast machines.  That mode could go
lots faster if they could make use of a device driver, since that
would allow video-card specific optimizations to be used (ie, you
shouldn't have to resort to boring svga when you have a $200 video
card).  In this instance, the lack of an OS makes the game slower
than it could be.

I've seen all these arguments before, and never has anyone backed
it up with anything concrete.

> Why is Microsoft
> wooing games developers so intensely with their "Direct Whatever"
> APIs that give programmers direct access to a virtualized hardware-
> level device ?

Because Microsoft Windows has historically been terrible at this
sort of thing.  That doesn't help or hurt your point, except if
you can't think beyond dos and windows as your only two choices.

> >1. Are inherently non-portable.
>
> Big deal. The PC is a hardware standard; programs should work on
> a PC independent of OS. What's non-portable is the tons of APIs
> and little supervisor services that stand in the way.

Um, in other words, we would have portability problems if
everyone used a PC.  Sheesh, if everyone stuck to bicycles,
we wouldn't have to worry about choosing between regular,
supreme, and super gasoline when we fill up.  And the PC is
not that standard, it's a shoddy architecture that barely works,
is incompatible with itself (try putting an OS more robust than
win3.1 on, and a lot of the hardware starts to fail).  The
tons of API's owes a lot ot Microsoft, the company that refuses
to use a standard unless they wrote it themself.

If you only use a PC, great.  Go off by yourself and stop
bitching about people trying to do more.

> >Of course, these points are non-issues for the MSDOS programmer.
>
> These are non-issues to personal computers.  And that's what DOS,
> Windows and OS/2 are supposed to be: personal computer OS's.

What about Macs?  Guess they don't fit in the picture, because it's
too hard to write portable code.  I think the only reason you put
OS2 in the list is because it has a DOS emulation subsystem.

> As for Unix, multi-user actually stays in people's ways; it helps
> nothing in a personal computer environment.

Not everyone uses computers the same way.  This means that not
everyone uses computers the same way you do.  Which means that if you
personally can't find a good use for something then you can't assume
that no one else can.  I personally, get more useful stuff done when
I log in on my X terminal to a UNIX box, than I ever do with that
piece of junk on the other table.  I don't think you would, because
you probably do different things than I do; but I don't go around
assuming that everyone needs a multi user environment, so why are
you going around assuming that no one needs one?
--
Darin Johnson
[e-mail]
    My shoes are too tight, and I have forgotten how to dance - Babylon 5