[MUD-Dev] Re: Modular MUD

["D. B. Brown" <dbrown1#stny,lrun.com>]

-----Original Message-----
From: Jeroen Ruigrok/Asmodai <asmodai#wxs,nl>
To: mud-dev#kanga,nu <mud-dev#kanga,nu>
Date: Sunday, August 30, 1998 11:16 AM
Subject: [MUD-Dev] Re: Modular MUD


>At 12:35 AM 8/29/98 , Ola Fosheim Grøstad wrote:
>>Caliban Tiresias Darklock wrote:
>>Well, I'm not too worried about Windows being buggy or about loose
pointers
>>or other bugrelated matters. What I worry about is how well it runs over a
>>long period of time. Microsoft is very much a "what does the average
>>potential buyer of this product demand" type of company, so I simply do
not
>>believe that they have focused enough on long term stability with
>>Windows95.   Maybe this is just my glasses being foggy or something, but I
>>have noticed a slowdown in performance at the end of longer sessions
>>(10hours+), which would suggest that Windows95 has a somewhat sloppy
>>cleaningup department.
>
>Aye, and NT suffers from the same fact, so I wonder if either Windows 9X or
>NT are suitable for long term MUD running, since it allocates, but doesn't
>clean up correctly.

I can attest to 95 working as at least a small-scale mudserver.  My mud in
development, Jormundgand, has been residing on a 95 box for the past year
and a half, and has suffered mayhaps two Win95-related crashes.  Win95
itself
has been brought down only a couple of times -- the only time in the past
few months was my fiancee's game of Diablo crashing it.  At present, the mud
has been up for nine days, the machine itself about two months.

>>Besides, do you expect the admin to dedicate a machine to the MUD, or
should
>>(s)he be able to say, run MSWord, at the same time?  If you expect the
user
>>to dedicate a machine to the task then you could simply let the user admin
>>the MUD from a remote client with a simple and neat GUI?
>
>Hmmm, I can almost taste the performance loss by running Word on the same
>machine... Windows isn't that great at utilizing/optimizing the
>swap/pagefile...
>Also, given the fact that Windows machines have proven to very lenient
>about the RFC specifications governing TCP/IP stack and POSIX compliance.


The performance loss due to Word is minimal, but somewhat more than I'd
expect
(To my dismay, all Office applications seem to suck down every availible
idle
CPU cycle, for reasons unknown to me).

But, overall, I've found that Win32 platforms are not horribly inferior to
Unices for running muds, (Note that I didn't say "not inferior", but instead
"not horribly inferior") contrary to general mailing list opinions.  I do
plan to move Jor to a Linux box in the immediate to near future, but it
certainly
isn't hurting on the 95 box it resides on now.

D. B. Brown

P.S. -- The only real problem I've discovered is in filenames, especially
the way MudOS interprets the relationship between mud objects and the
file which spawned them.  Win32 thinks that "/U/VII/FOO" under MudOS is the
same as "/u/vii/foo", however, MudOS thinks that they are seperate... thus
one can load each as a seperate object.  More of a cause of confusion than
anything else, but still a problem.