MOO-cows Mailing List Archive

[Subject Prev][Subject Next][Thread Prev][Thread Next][Subject Index][Thread Index]

Re: moving things out of the server (was: floats and efficiency)

At 10:47 AM 7/22/97 +0300, Gustavo Glusman wrote:
>Seth writes:
>>... and while we're at it, can we finally move .location and .contents
>>(and other spatial assumptions) out of the server?  This could be done
>>in-DB, with the server's parser declining to work is there doesn't exist
>>properly-formed .location and .contents properties where `player' is.
>Similarly, why not do without the 'name' property on every object?
>I am not suggesting this. What I mean to say is, that just as you don't
>have to use floats just because they're there, you don't have to use the
>location/contents system just because it's there. But I think most MOOs out
>there actually use it.

What I've always liked about the MOO server is that it's fairly minimal.
There are very few commands directly supported by the server, and even
its builtin parser can be dispensed with, if needed.  This makes it very
suited for database work of many styles, since the language supports so
much potential, and the server doesn't impose much of a specific model 
on it.

However, the model which the database -does- impose includes spatial
assumptions.  Not all -- perhaps many, but not all -- database applications
have use of location or contents information.  For those apps which don't
need this information, it's wasted space.

It would not be the most difficult thing in the world conceptually to move
location and contents into the database (perhaps as a compile-time option)
so that wizards can code their own location system if desired (3D? whatever)
or omit it entirely.

Seth I. Rich -            Rabbits on walls, no problem.

Home | Subject Index | Thread Index