Re: [wcsplus] explanation of wcsplus initiative

I totally agree ... Complexity belongs with the small group responsible
for figuring out the abstract concepts; we MUST give the implementers /
service consumers something EASY to use or they will look elsewhere!

Jeremy


-----Original Message-----
From: wcsplus-bounces@xxxxxxxxxxxxxxxx
[mailto:wcsplus-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Stefano Nativi
Sent: 05 October 2007 10:14
To: Ben Domenico; wcsplus@xxxxxxxxxxxxxxxx
Subject: Re: [wcsplus] explanation of wcsplus initiative

Hi Ben,

I agree with you on the need to make it clear that this is an effort to
"developing a specification that will be much different from WCS 1.1."

To come up with a clear statement sounds very good by me. In your
message you hit main points, in my opinion.

In the same line, I'd like to stress a useful approach: work on complex
abstract specifications to keep implementation simple. In fact, the
Coverage realm may be very complex (i.e. the reality is complex! [John
Caron]). Hence, this complexity can't be avoided, but it should be
addressed and analyzed only at the abstract level (avoiding shortcoming
and allowing extensibility), but we must not shift this complexity at
the implementation level. This is possible by making precise (and
strategic) implementation choices. Actually, Abstract specs are for
domain experts, while implementation specs are for developers and users.

Perhaps this is obvious, but, in some context, it is difficult to make
precise implementation choices (e.g. multi-community and heterogeneous
groups) and the abstract complexity is just encoded in the
implementation spec.

In my opinion, WCS 1.0+ should try to capture the key implementation
choices for FES community looking at interoperability with other
communities (i.e. the compliance with the full abstract spec).

I think this is useful to the WCS RWG activity, indeed.


--Stefano


_______________________________________________
wcsplus mailing list
wcsplus@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:
http://www.unidata.ucar.edu/mailing_lists/=20