Re: xml processor profiles diagram

"vojtech.toman@emc.com" <vojtech.toman@emc.com> writes:
> Great job, it looks very nice - ...and quite complicated at the same time.
>
> I think the reason is that the diagram tries to combine three types of
> things together: not only the profiles and the classes of information,
> but also "uncategorized" (the two yellow sets) processor properties
> such as base URI processing and xml:id processing. I think that having
> three different types of things in the diagram makes it rather
> difficult to understand.
>
> I am actually wondering whether the fact that the diagram has to rely
> on additional "uncategorized" properties rather than just on the
> profiles and classes isn't a bug in our specification.

I'm having a hard time seeing it that way. The xml:id circle shows
which profiles appeal to the processing described by an additional
specification. We'd have factored xml:base processing out in a similar
way, I suspect, if we'd been willing to consider a profile that
doesn't mandate it.

> I think that the outer set (processing as required by conformant XML
> processors + base URI processing) is actually a subset of the Core
> class and can be removed from the diagram.

The outer-most circle could be removed, I suppose, since it's
universal over our profiles. But I don't find it confusing.

> However, there is no class that would correspond to the other yelow
> set (xml:id). Should we have one?

The property of being an "ID" requires the Extended and Decl classes,
I think. I don't think we need a new class, but I could be confused.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh
Lead Engineer
MarkLogic Corporation
Phone: +1 413 624 6676
www.marklogic.com