Re: CF: Proposal for features for a random map generator

[Peter Mardahl <peterm@langmuir.EECS.Berkeley.EDU>]

> On Dec 10,  4:44pm, Peter Mardahl wrote:
> > Subject: CF: Proposal for features for a random map generator

> probably be some information to the player what maps are random and what are
> quest maps (more clearly defining the quests is probably worthwhile).

Yes.  I'd planned to mark each randomly generated map with a sign at the
arrival point in it.

>  Keeping the maps the same each time - you probably want to keep the walls th
> same, but not necessarily the monsters (otherwise you run the risk of some ma
> by pure luck having a great item on it, and players repeatedly visiting it
> knowing they can get it.  There have been a few times where I got some great
> item out of a pretty low level map - rare, but possible).  Also, it would be

I don't anticipate that the items would be the same, because the artifacts
code would have a different random generator?  Maybe it would though.  Hmm.
Well, maybe it wouldn't, since the random
map would be set up by a standalone program, and would have an independent
seed.  Artifacts wouldn't get made until the server "loads" the map.

I'll make this a config option I guess.

> >   Ubiquity:

>  I think that any exit that has some flag/parameters set that states it goes 
> a random map should do so.   There are probably numerous broken exits right n

OK.  I can make this a config option, too.  Can you think of a way to make
random maps obviously different from normal maps?  So that every building
need not be searched?

Or should we simply require that random maps be turned on exit by exit?

Or how about this:  the parent map must have a flag set that its exits
lead to random maps?  I think this is the best idea.

> >   Quests:
> >    It would be cool if random quests could be implemented, too.  I won't
> > put this in immediately.

>  I would think that the ability of quests would be having an item/reward at t
> end of the puzzle and somehow distributing the information the maps.  This ma

Yes.  This would be pretty tough to do well randomly.  

> >   Standalone operation:

>  Yes - ideally, it would be a function call like make_random_map with various
> parameters passed - the map it makes could then be dumped to a file, or kept 
> memory in case of maps needed on demand.

I was actually thinking of having crossfire call it using a system call, and
then loading the map.

> >
> >   Size:
> >    Random maps should be sized.  A big, multi-square exit should have a
> bigger
> > random map to it, while a small house should have a small interior.
> >
> >   Inheritance:
> >    The type and location of the exit should imply something about its rando
>m
> > interior.  A shop exit should have a shop interior, a house should have a
> > home interior, etc, a difficult map should have difficult sub-maps.
> 
>  I think this should be set in the exit structure - if we go by the theory th
> the only random maps are the actually specified, than this is not a problem.
>  Shops are 4 space objects, but actually not very big inside (many houses are
> bigger).  So image size of the entrance may not be the best method.

How about doing this only in the case where a specification is not made?

Anyway, any important features that I've missed?  Volunteers to help?

PeterM



-
[you can put yourself on the announcement list only or unsubscribe altogether
by sending an email stating your wishes to crossfire-request@ifi.uio.no]