Paul wrote:
>Don't get me wrong, I'm not suggesting a crosslink is a bad thing, nor
>that flexibility isn't necessary, but I believe it would be more useful
>if you _eventually_ specify roles better than just role is any string,
>i.e. CDATA.
Useful to who? One of the reasons I like to leave things unconstrained
is because the variety of roles that can be created is so prolific. Good
software design doesn't force the user into a predetermined
configuration but guides them to set their own "local" configuration.
For instance, suppose we specified all the important roles that belong
in a cave survey recorded in CaveXML, and left an ability to add new
ones as you suggest below. Guess who's NOT going to find that useful,
all the French, German, Japanese, etc. cavers who don't want to use
English to describe the roles in their survey party. You'll just annoy
many of them by demonstrating that all too common Anglo-centric view of
the world (and yes I'm an American of English, German, Irish decent).
>role=bookie, role=sketcher and role=book all look fine for many humans,
>but I believe it is more useful when the values are constrained to
>define values that connect each instrument (probably just use the
>same names as the instrument types). Additional free-form values
>would also be allowed so that you wouldn't be limited to the recording
>the known roles.
Free form values would be mandatory.
>Now it is up to everyone who is defining this particular PSC CaveXML to
>define when 'eventually' is.
Aahh, you're helping me do that now. This forum has been a wealth of
information on how one might or might not go about representing cave
survey data in XML. Thrown up against my systems design background I've
learned some valuable lessons from this group. A lot of them on the
subject of scope control. I've just reconfirmed for myself that it's not
the technolgoy that's difficult, XML is really simple, it's the meeting
of the minds.
>Have fun.
>-Paul
Yes it has been.
-- Devin Kouts Caver Systems Engineer www.psc-cavers.org
This archive was generated by hypermail 2b30 : Mon Apr 02 2001 - 18:00:00 CEST