[MUD-Dev] Looking for books...

[coder at ibm.net]

On 19/08/97 at 09:55 PM, Ola Fosheim Gr stad <olag at ifi.uio.no> said:
>Greg Munt wrote:
>> 
>> I know some of you have some reasonably-sized libraries, so I was
>> wondering if you might be able to recommend some books for me?
...
>>    Database design (esp persistence)

>Perhaps you would be better off by using a library for persistence?

Texas, Yooda, tdBm with extensions, and various others as listed on
the Free DB list come to mind.

>Do you really need any advanced database design in a mud? Do you
>really need real persistence, or is saving the most vital attributes
>enough? (ought to be)  Persistence isnt always very cool if your code
>has bugs (depends on the approach I guess).

Many, not all, of us here are discussing creating full worlds with no
concept of object or world resets.  As such persistance of objects,
and object changes are implicit to the world design.  The base
assumption thus changes from "save what we have to" to "everything is
saved by default, explicitly not-save what we don't want".  

>>    Tree algorithms (esp R-Trees, et al) (all I can find are general
>>    algorithm books that discuss binary trees - sigh)

>What is an R-tree, why bother with more complex things than binary
>trees (or a self-balancing tree (splay?)) and hashtables if those are
>sufficient?

Because when you move from a room based model to a fully coordinate
based model your data set grows exponentially.  A significant problem
then becomes determining what the working ste of objects for a
particular event is with the least amount of computational overhead. 
Various tree form data structures, such as R-Trees, R*-Trees, V-Trees,
Oct-trees, K-trees etc offer various perfomance and manipulation
returns here which can then be selected from for best performance on
*your* loading.

I'm particularly fond of the Stony Brook book for this area as a
fairly thorough survey of the various data structures.

>>    Parser/NLP design

>What is NLP?  

Natural Language Processing.

>The compiler book covers parser design, though for a
>startup project a simple approach would be sufficient, dont you
>think?

A lot of us are a long ways beyond startups.  We also have several
contributors from well known MUD server bases, and from commercial
offerings like Ultima Online, the TEN Network, etc.  The idea is to
extend the state of the art by at least an order of magnitude of more. 


My own suspicion is that real NLP is something to be avoided in a MUD
server.  Reducing your language to a base set of expression forms (ala
the good old standby: verb noun) and then requiring that more minimal
set of constructs (which can be very flexible as LIMA and Island show)
is the best route out.

>MOO was
>written as a research project and the author is with Xerox Parc... 

MOO was written by Stephen White, and was some of the worst code I've
seen.  He then rewrote MOO and set it up at Xerox PARC with Pavel
Curtis.  Stephen then became somewhat disgusted with the still poor
design of MOO, and handed it off to Pavel who renamed it LambdaMOO
after the site it ran at,  Stephen then wrote COOLMud, and absolutely
incredibly elegant and tiny pice of code which should stand as an
example to any MUD server author.

Hurm.  I guess this is a bit of a fractious welcome to the list. 
Sorry about that.  Welcome!  

--
J C Lawrence                               Internet: claw at null.net
----------(*)                              Internet: coder at ibm.net
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...