On Fri, 22 Aug 2003, Marshall Rose wrote:
great. if your confident with an alternative approach, put an
initial document together and submit it. i don't want to forestall
alternatives, but we need to be wrapping things up.
An alternative draft is attached in plain text. HTML rendering is at
http://www.measurement-factory.com/tmp/opes/snapshots/current/p.html
As we discussed, this draft is not as detailed or as polished as the
IRML draft. The draft purpose is to illustrate the "killer idea"
behind the alternative approach: a minimalistic and efficient
configuration language designed specifically for the problem domain,
but with a room to grow. See draft introduction for more detailed
treatment of design goals.
I hope this draft is sufficient to pick one direction over the other,
but I would be happy to clarify or add as needed, of course. I do
realize that there has been a lot more work put into IRML already. I
wish I knew about its existence earlier. If nothing else, this draft
is probably the ultimate form of "constructive comments" IRML folks
are asking for :-).
On the positive side, many IRML ideas/developments can be easily
reused if an alternative approach is chosen. The reverse is probably
not as true because P capitalizes on being a language rather than a
hierarchical collection of attributes (but it is all relative and
equivalent from Turing point of view, of course).
HTH,
Alex.
P.S. If "P" is already taken as a language name, please let me know.
My quick search yielded no matches. "PL" does not count, IMHO.
p.txt
Description: draft-rousskov-opes-rules.txt