Re: JSR-000093: JAXR - your opinions?

[Nikola Stojanovic <nhomest1@twcny.rr.com>]

<Bob Sutor>
although possibly with less inside knowledge of ambiguity resolution that is
in the heads of the implementors.
</Bob Sutor>

That lack of inside knowledge should help in providing the feedback to spec
creators so spec could be refined and made less ambiguous.

<Bob Sutor>
ebXML will have to deal with 4) and 5) eventually.
</Bob Sutor>

Section 10 of TA spec talks about ebXML Conformance and Conformance testing.

<Bob Sutor>
In the same way, language-specific bindings can be done by the spec
implementors or by another organization.
</Bob Sutor>

And by both (many) together?

<Bob Sutor>
Perhaps the spec implementors can do a better job at the bindings, if they
have the resources available. This
also makes the general standards/binding development process more scalable
than always doing in one organization associated with the language.
</Bob Sutor>

I can see your point. If the binding is 1 to 1 then it might make sense. In
fact, it looks like that (so many times mentioned) ODMG (OQL) group did
exactly that. However, if you want to do Many to 1 then you probably cannot
stay with only 1 spec creator. There was recently a post to this listserv
http://lists.ebxml.org/archives/ebxml-regrep/200101/msg00127.html that
refers to "OneBasketPattern". I hoped that ebXML is all about
"OneBasketPattern", but have realized that some patterns could be
implemented in "AntiPattern" fashion as well.

Regards,
Nikola