Martin Bryan wrote:
> Eric
>
>
>>Being streamable is what enables applications such as schema guided
>>editing as demonstrated by James in Philadelphia and I think we should
>>keep that requirement for most of the parts (including part 4).
I disagree that streamability should be a requirement: may I suggest
a weaker substitute? That each spec that has a non-streamable default
also provide a mechanism/option for a streamable version. That is what
Schematron and my proposal for Part 7 do.
Things like ID checking are not streamable, in a sense, by the way.
What happens in editor implementations (such as James' and mine)
that do not build a full DOM is that you elect not to continuously
report certain constraints. In other words, you end up with a kind
of progressive validation anyway.
That being said, I personally would be very happy if there were
a maximal streaming subset of XPath standardized: I am thinking of
STX that Christian Netwich among others have been working on.
I would be happy to include bindings to that in ISO Schematron
and Part 7 Charrep.
> Whilst editing is a classic case of why you want partial/streamable
> validation, it clashes with the need to do "overall validation" prior to
> archiving results. We need to keep both applications in mind. For example, I
> may need to be able to check that at least four items have been ordered in
> the complete order, or whether a charge has been made for transportation
> costs iff the total order value does not exceed a particular value. These
> sort of checks require completion of the order prior to validation and are
> not necessarily streamable.
+1
Cheers
Rick Jelliffe
-- DSDL members discussion list To unsubscribe, please send a message with the command "unsubscribe" to dsdl-discuss-request@dsdl.org (mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)Received on Thu Apr 8 12:54:54 2004
This archive was generated by hypermail 2.1.8 : Fri Dec 03 2004 - 14:00:28 UTC