AC2 was RE: [MUD-Dev] Total Annilation of Downtime

Brian Hook brianhook at pyrogon.com
Mon Dec 16 16:41:05 CET 2002


[Daniel Harman]

> Or why not just treat heightmaps as a specialised mesh in your
> general engine?

I'm sure most of them do in the backend graphics pipeline, but the
fundamental problems I outlined in another e-mail still remain --
multiple parallel content creation toolsets, and anything that takes
advantage of the special case nature of the heightfield will, of
course, be special cased for the heightfield.

For example, with most regular meshes, you'll have manually
specified texture coordinates.  With a heightfield, you will almost
always want to have procedurally generated texture coordinates to
save on space.  Heightfields will also have implicit XY, and likely
will have 16-bit integer Z values (assuming Z = elevation).  The
differences add up.

> What looks to be a bigger problem is that developers have removed
> zones, or have at least implemented invisible transitions. They
> then have a really limited working set of textures+meshes for the
> whole world. I guess limiting the potentially visible set with a
> zoneless world is actually a pretty tricky problem, especially if
> you don't want people fleeing monsters to be able to reach a
> certain point past which a monster can't chase them further.  I
> imagine being able to step over an invisible line and see a
> monster wink out of existence would look a little silly.

I think what most developers will do is have a monster on a "leash",
and the "pull" on that leash is probably a fuzzy combination of
morale, anger, distance to player, distance to home and possibly
duration of pursuit.

Throw into a big equation with appropriate co-efficients for a
specific mob, and there you have it.  A D&D style "invisible
stalker" basically never gives up.  A giant ant, however, will
probably give up if it gets too far from home.  Etc. etc.
 
> Personally, I find a low poly world with good textures more
> interesting than a high poly world with poor texturing (which is
> why I think the PS2 sucks).

Well, that's kinda what texture mapping was invented for =)

> also because simplicity often looks nicer than the fussyness of
> newer games models.

As my partner and I like to say, "good art trumps good tech".  The
best engine in the world will always be held back by poor art, but
great art can usually still look good even in a mediocre engine
(e.g. Icewind Dale 2 or Baldur's Gate 2 -- extremely mediocre
engines, but gorgeous prerendered art)
 
> Maybe I'm just jaded, but it seems that the majority of these
> products fail on the small, but incredibly important details that
> would probably have taken 2 weeks to sort out prior to release.

I completely agree.  From what I've seen, almost every single one of
the major development houses has completely and utterly failed in
the departments of the "small things", proper coverage testing, etc.
As much as I like AC2, the fact that different types of text didn't
show up in different colors in your chat window -- I mean, come on.

Of course, with such a huge list of things to manage, 100% coverage
is tough to get through, and it's easy to fixate on the screwups
that manage to make through.  It's like being a sysadmin -- no one
remembers or cares about the 99.99% of stuff you got right, but
they'll forever remember when something broke.

-Hook


_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev



More information about the mud-dev-archive mailing list