[MUD-Dev] size

Daniel.Harman at barclayscapital.com Daniel.Harman at barclayscapital.com
Mon Oct 6 10:51:16 New Zealand Daylight Time 2003

From: Amanda Walker [mailto:amanda at alfar.com]
> On Thursday, October 2, 2003, at 11:07 PM, Bo Zimmerman wrote:
>> From: Amanda Walker
>>> Daniel.Harman at barclayscapital.com writes:
>>>> From: Amanda Walker [mailto:amanda at alfar.com]

>>> If you're not paying license fees for your back story or
>>> setting, don't need to use absolutely every feature of the
>>> latest Video Force Turbo 9000 video card, etc., why roll your
>>> own engine?

>> Because you can do it better, faster, smarter.  Because the
>> legacy engine is flat and difficult to extend.  Because you have
>> new ideas that don't fit the legacy engine.

> Sigh.  If we go back to the *context* of this discussion, the
> claim was being made that graphical MUDs take a lot of money and
> time to develop, which is why they require a big bankroll.  Sure,
> if we assume they don't, that does solve the problem, but that
> wasn't really the question at hand :-).

Well I'd argue that the 3d engine is the least of your problems
anyway. Its generating the meshes, the the textures, the music, the
sounds, the animations.

Until recently most buy in engines were woefully inadequate for
outdoor terrain anyway. Even now, very few are suited to zoneless
spooling in of data*.

Beyond those issues, 3d games require considerably more bandwitdth
than a text mud (in the order of 1k a sec is a typical target),
which makes it hard to justify if you aren't billing. The server
code is also by necessity considerably more complex as you can't
make simple assumptions about rooms, but instead have to have
complicated spatial partitioning routines to limit working set
size. Hell even writing path finding for a 3d game is a monumental
task compared to in a mud.

I haven't written my own UDP library just for the hell of it,
nothing out there suited my specific needs. If I could find
licenseless stuff that did, I'd use it gladly (and yes I have looked
at enet).


* Not that I think zonelessness is a good thing given the current
  limits on texture memory - massive height mapped terrains that all
  looks the same are utterly pointless.
MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list