OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

ubl-ndrsc message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]


Subject: [ubl-ndrsc] ur-schema necessitates xsi:type everywhere in instance


Title: Message
One issue with ur-proposal that was not discussed was this issue (from a private message I produced on Monday -- also, see the attachment):
 

... I worked out a little example of the ur-type proposal a while back. I think I've forwarded it before, but to save you groveling about the list I've attached it again.

I've made a namespace for ur-types and a dervied one for "UBL types". Then I made a third for a "third-party-restriction".

A key issue we identified with Paella is as far as I can tell, still an issue with the latest ur-type proposal. If you look at a "vanilla" UBL instance document (ubl-doc.xml) you see it has to use xsi:type on every element. The same holds for a "customized" UBL instance document (third-party-restriction-doc.xml).

The reason this happens is that every element declaration must be of an ur-type -- never a (derived) UBL type, or a (derived) customized type. It must be so since this is the means by which we are able, after the fact, to derive from the ur-type without the horrendous "cascade" back through all the referenceing content models (and the ones that reference those, and so on -- remember Barcelona).

The fact that the instance documents must carry xsi:type everywhere means that schema validation is _not_ (on its own) enforcing a particular vocabulary -- I can really use an ur-type wherever I want to and the schema validator isn't going to squawk.

If that's the case, then what, exactly is the value of the UBL namespace (i.e. the non-ur-types)? I mean, should we just ship the "everything is optional" ur-types and be done?

Maybe I misunderstand the proposal. If so, could someone please correct my simple example and show me some new instance documents that don't require xsi:type everywhere? Barring that we should probably revisit our discussion from last March and decide whether that's ok, or whether we need to fix it somehow (and no, I don't have any ideas just now :-)

One thing I have tried is to get the xsi:type attributes on the elements in ubl.xsd to _default_ to the UBL types somehow... But I fear I am not nearly the XSD wizard for that task -- heck I can hardly even describe it.

 

Attachment: ur-example.zip
Description: Binary data



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]


Powered by eList eXpress LLC