|
Martin,
Thanks for forwarding Chris' notes. First
item - I checked his membership - and
will look for a match on that sender.
I really like the idea of CAM-base, CAM-patterns,
CAM-repositories as the three
components.
And then the points on schema-compatiblity are well
noted. As Chris' notes - the
recursive functionality is intended to replicate a
structure, not insert it arbitarily -
(we have insertTree() to cover those
needs).
We have also received many good comments from the
OASIS review period.
What I'd like to do next is consolidate all this
into a new draft - rather than rush
out a V1.0 - and make sure we cover all these good
suggestions. I've just not
have time to get to this with work pressures - and
looking right now at my
schedule - its going to be July before I can
realistically dedicate solid time to
it. Since I've also noticed that many of our
examples in the spec' are now
deprecated by the latest schema - its quite a task
to get this all done.
Anyway - that's my sense for now - that we can work
over the summer,
completing the V1.0 details and then have that
ready late August to
re-present to OASIS formally. This also gives
us time to work with some
more implementations and really validate V1.0 some
more.
BTW - ITR - working with financial transactions
that I'm doing right now I just
uncovered the need for basic date functions - today() and simple
date
comparison +1D, -7D, +1M, +5Y, and so on - to augment the XPath
comparisons we're using. I'll write something
up in email on that later
this week.
Thanks, DW
----- Original Message -----
Sent: Monday, May 17, 2004 4:36 AM
Subject: [cam] FW: CAM Thoughts
From: Hipson,C,Chris,XSJ61 HIPSONC R
Sent: 13 May 2004 15:16
To: Roberts,MME,Martin,XSG3
R
Subject: CAM Thoughts
Hi
Martin
Having discovered that I can’t
post to the CAM list [bad attendance record I
guess J] I thought I’d run a few
thoughts by you.
Overall CAM looks nice although I imagine
it won’t fly well with existing schema fanatics J [not
complex enough].
I like the way that CAM
provides ‘deterministic flexibility’ sidestepping the XML-Schema issue that
a flexible schema makes everything optional.
However assuming that users will start from existing
XML there appear to be some omissions?
1)
There is no
predicate support for namespaces. [Yes I know you don’t like them but
if you need to meet existing XML documents half way then you’re going to
need this?]
2)
In XML-Schema
xs:all allows a ‘group’ of elements to appear in any order rather then in
the order listed. While I doubt there’s much demand for this in business
documents it would be very hard to reproduce in CAM
without explicit support? So may be worth adding?
makeRecursive
Requires a StructureID which is only available as an
Advanced Option? But the predicate itself is not itself an Advanced
option?
“Donates that the specified
parent element can occur recursively as a child of this
parent.”
Does this mean that a block of XML could appear
anywhere in the tree under the
parent?
This would seem excessively liberal and would
prevent any hope of building an XML-Schema from a CAM
template as David Webber would like to do?
Surely you need to have some ‘link’ from the point
the recursion is ‘inserted’ to the structure that is to be
‘recursed’?
Or have a missed
something?????????
Packaging – trivial observation but the current
vogue [at least in web services J]
is to try and make large specs appear smaller by chopping them
up!
In the case of CAM
this might fly quite nicely as CAM Base would equate to your stuff and
CAM Patterns and
CAM Repositories to
Davids?
How’s the selling process going with UBL?
Where does it sit with respect to ebXML [which it obviously sits quite well
with]? And are you looking anywhere else? ISO have for instance DSDL
which is a schema collective wrapped round XML Schema, RELAX NG, Schematron
etc. etc. – might offer some scope for feedback if nothing
else?
Cheers Chris
Hipson
Web Service Technology
Consultant
"If the facts don't fit the theory, change the
facts." Albert Einstein
Tel. 01977
593109
|