Re: OPES Rules Language

On Wed, 2005/06/08 (MDT), <[email protected]> wrote:

> I'm going to give the rules language a shot by trying to author
> a draft that merges the best of P and Sieve.  Simple semantics
> and rich matching,

Will you extend Sieve? In other words, will your language be a
Sieve extenstion, following Sieve extension requirements?

Will you make your language specific to HTTP adaptations?
Or will it have modules/libraries/extensions/whatever for
each application protocol being adapted (HTTP, SMTP, etc.)?

> with control points tailored to the OPES architecture.

Do you mean points where adaptation is triggered? For
example pre/post-cache and request/response path in
caching HTTP proxies? If not, please ignore the paragraph
below.

In my limited experience with ICAP configurations, and based
on the "application agnostic core" principle, there may be
no need for explicit control points. The control point can
(should?) be an external to the language matter. The language
does not have to depend on the message processing "point"
where the adaptation is triggered. And IIRC, OPES architecture
only provides examples of possible control points.

> Send your wish lists.

My wish list can be easily derived from the Sieve versus P
comparison I have posted some time ago. I want all the good
features that made the basis for that comparison :-).


Thank you,

Alex.