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


Help: OASIS Mailing Lists Help | MarkMail Help

security-services message

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

Subject: RFE related to xsi:type on AttributeValue

The issue I raised at the end of the call was a request from Steven Legg to
consider relaxing the rules on xsi:type in core that dictate that if
xsi:type is used on an AttributeValue element, all value elements of that
Attribute share the same xsi:type. So none, or uniform, basically.

He thinks this is too limiting because he has a use case in which the value
type of an attribute is part of an XML-defined hierarchy and a given
attribute can contain values of a base type or a derived type and he'd like
to signal that at runtime using xsi:type.

I'd have to say it's an RFE, not an errata. The language in core is really
best read as a suggestion that xsi:type has so much baggage that people are
better off creating their own signaling mechanisms in their profiles.

Maybe we should just say that. That's probably good advice since QName
values are so evil in every way. We're stuck using xsi:type in various
abstract extension points, but the less used the better.

-- Scott

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