This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Comment from Francois Richard, only entered into bugzilla by Felix: Currently, an ITS implementation has to decide whether it wants to use in situ or dislocated scope / selectors. Would it be possible to use only dislocated scope / selectors, and in situ only without selectors (but with a default scope for each data category)? That would make the conflict resolution between in situ and dislocated easier, and the overall design of ITS, without loosing expressive power of selectors (since every in situ selector can be expressed dislocated as well).
Very good comment/idea which also occurred to me. From my understanding, it could give rise to a set of data categories which is related to ITS itself. From my point of view, this set looks somewhat like 'elementFormDefault', 'attributeFormDefault' etc. in XSD. To be specific: We could think about a data category like 'Selector Type' which for example could be coded as an attribute 'selectorType' with the possible values 'in-situ, dislocated, both, unknown'.
(In reply to comment #1) > Very good comment/idea which also occurred to me. > > From my understanding, it could give rise to a set of data categories which is > related to ITS itself. From my point of view, this set looks somewhat > like 'elementFormDefault', 'attributeFormDefault' etc. in XSD. > > To be specific: We could think about a data category like 'Selector Type' which > for example could be coded as an attribute 'selectorType' with the possible > values 'in-situ, dislocated, both, unknown'. I don't like the idea of putting a selection related aspect into the data categories. And no matter if we call this "data category" or not: 'selector type' makes possibly things complex. What should happen, if the type is "in-situ", but there is dislocated ITS information in the document? I think the initial idea of Francois's comment was to make things easier, not more complex. I think this question is more one of conformance to ITS: does an implementation support ITS information in a schema, and / or in situ and / or dislocated? If we want to allow these variations, we could specify each of the three as an optional implemenation feature, which would be the same effect as with the attribute values 'in-situ, dislocated, both, unknown'.
The working group decided to adopt this proposal. See http://lists.w3.org/Archives/Public/public-i18n-its/2006JanMar/0098.html