[MUD-Dev] Text Parsing

Chris Gray cg at ami-cg.GraySage.Edmonton.AB.CA
Wed Jun 2 21:44:56 CEST 1999


[Cynbe ru Taren:]

> What -users- want is a simple, compact notation with easy to
> understand meaning.  (Lots of conflicts and trade-offs hidden in that
> spec, of course!)

Agreed. However, for the new crop of non-computer-savvy users, making
that input as close as possible to simple pidgin English (or whatever),
is likely to minimize their frustration.

> What attempts to do "natural language parsing" usually produce are the
> opposite: complex, verbose, inconsistent notations where it is
> anybody's guess what the machine will do in response to a given input.
> Normal users find this very frustrating.  Predictability is -very-
> high on the list of what users want from a computer interface.

I definitely agree with wanting predictability, but, depending on just
what the definition of "natural language parsing" is, there is no
necessary connection between them. As I mentioned in my previous post,
one big source of unpredictability is arbitrary abbreviations (or ones
which change over time). When I'm picking something up in my MUD, I
don't type 'Pick up the big red book.', although I could. I usually
type 'get book', or, if I remember what abbreviation I put in for 'get',
I'll use it. However, unless the system is told what all words are
adjectives, and no words can be both adjective and noun, then I can't
just type 'get book bag', but must use 'get book from bag', simply in
order to make the command easily interpretable. If you have complex
rules for trying to figure out, based on world context or previous
commands, what 'get book bag' actually means, then you have lost the
predictability of the interface. Also, without the prepositions, you
are limited in what forms of verbs you can support, and you are back
to forcing the user into the "guess the verb" game.

I agree with most of the rest of Cynbe's note, although I was not aware
that there was a formal 'Algol' before the original 'Algol'
programming language. Got any references?

--
Don't design inefficiency in - it'll happen in the implementation.

Chris Gray     cg at ami-cg.GraySage.Edmonton.AB.CA
               http://www.GraySage.Edmonton.AB.CA/cg/


_______________________________________________
MUD-Dev maillist  -  MUD-Dev at kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev




More information about the mud-dev-archive mailing list