Mail Archives: opendos/1997/05/19/21:02:08
> > To all those thinking about X on OpenDOS, *PLEASE* go to
> > http://ecco.bsee.swin.edu.au/unix/uh/x-windows.html and read it ALL. If at
> > the end of this you're still in the mood to have X in OpenDOS, take 2
> > aspirins and go to bed. You can call in sick for work.
>
> I agree with everything this document said, having tried
> most of it at least once. In my opinion, one reason
> Unix is on the decline is the despicable nature
> of X-Windows. I find it rather humorous that people
> who hate Windows and love Unix are even willing to
> say anything about X-Windows. There are reasons
> few programs are ported from Windows to X, and
> most of those are in that document.
The "X-Windows Disaster" document (chapter 7 out of the UNIX-HATERS
Handbook) is biased (to say the least) and out of date. There are some
good points in it, but I can say that X is a very good idea, and most
Linux users will agree it is great. But I must add that it isn't
necessarily *appropriate* for everything.
Here's my experience: We have a mixture of PC's and Sun (68k & SPARC)
equipment here; the beauty of X is that we can run old PC's (like 386's
and slow 486's) perfectly well as X terminals when they are far too
useless to run MS Windows and the latest bloatware. I am typing this
on a 1987 Sun 3, which is perfectly good for running the powerful
editors, compilers, maths tools, e-mail etc that I want. Compare this
with those running MS-this and MS-that who have to upgrade hardware and
software every couple of years... not just a big cost but a
disruption.
The security model is great. Again: compare this with the equivalent
in the MS-Windows world (just recently we had a demonstration of how to
"correctly" install W95 and NT systems, and they went to great pains to
edit the registry, store student-dependent config & personal files on a
Netware server (Netware 'cause we all know the MS equivalent isn't safe
enough). Fine. Now watch as somebody comes along and supplies an
explicit server name at the login, so they use a server where they have
less stringent restrictions, and the workstation is open to abuse!
I have *much* less problems supporting 50+ X11 users than people in
other departments here with traditional PC setups. If a user trashes
his setup files it only effects one person, and even then it is a
trivial matter to copy the handful of "standard" ones back for them.
The only problems with differences between systems are questions like
"where is the COMPOSE key on this keyboard" and a few odd fonts an
application might prefer. I can also add new, interesting software to
the system (and upgrade it) much easier, and know that if it works for
me it will work for others.
Some years ago, X11 needed what was a big computer at the time (when
2Mb would run your MS-Windows stuff very nicely). Now it is *easy* to
run networked X-Windows (without swapping) and (say) a 640Kb DOS
session all on an 8Mb 486 33MHz that wouldn't even run W95 plus
networking well, let alone Office 97. There is a smallish cost in
updating the remote hosts in a situation like ours (not really so in
most home-PC Linux setups, but that is because we have some really huge
applications - and when we talk about a huge application it might be
Mathematica or Macsyma or something like that which really does need
over a 100Mb of RAM for an array they mustn't get swaped out!) It is
rediculous to try to upgrade all desktop PC's so they can run the
biggest apps if they ever need to. That big-RAM requirement might have
sounded foreign to average office-app users just a few years ago, but
look at the requirements for modern word processing and other "core
user" work today!.
Now, X11 isn't really good on 640Kb 286's, let alone some of the
bottom-end equipment that plain DOS and even ViewMax can run on. But
the approach is a good one. X11 has tons of features that not everybody
needs, but it could be possible to adopt enough of it so a program
develped for XMax or whatever also can run with X11, and
administration/learning time is shared.
I think what is needed is not just a GUI like X11 but a program
developement system to allow everybody from novices to experts create a
wide range of software. To begin to decide on something like that we
have to look really hard at what it is that people who use systems like
X11, OS/2's object-oriented GUI, NeXTSTEP, etc appreciate, and then
look at what is going to be doable and popular in a MS-led world.
I'll post my thoughts in another message, so you've got time to think
of your answer!
-------------------------------------------------------------------------------
Mark Aitchison, Physics & Astronomy \_ Phone : +64 3 3642-947 a.h. 3371-225
University of Canterbury, </ Fax : +64 3 3642-469 or 3642-999
Christchurch, New Zealand. /) E-mail: phys169 AT csc DOT canterbury DOT ac DOT nz
#include <disclaimer.std> (/' "42. Or, on an early 586: 41.9999986"
-------------------------------------------------------------------------------
- Raw text -