[xml] Creating a markup language

[Bob Rossi <bob brasko net>]

Hi,

This is the best place I could think of for asking this question. I hope
someone is capable of providing me with an answer or maybe pointing me
in the correct direction.

I am creating a protocol between two GPL'd applications. Basically,
process A sends a command to process B. Which then responds with the
reply. I want to make process B respond in XML. This way, process A
doesn't really have to worry to much about parsing.

Basically, my question is, should I create a Markup Language for B that
it will always reply to A with? Or, should I create one Markup Language
for each command possible that B responds to A with? How can you
compare/contrast these two models?

If there were ten possible commands A could run through B, then the first case
would have 1 large ML, while the second case would have 10 small ML's.
I don't know which is better?, how they scale?, or what path I should go
down. I can say that the output of B to A is generally small.

Thanks for any help or advice,
Bob Rossi