Cart
[Home] [By Thread] [By Date] [Recent Entries]
>One area where I'm still very ambivalent is whether different messages >should use the same namespace, given that they might apply different rules >to the same elements Interesting. Some suggest that it is absolutely essential that each data item that appears in the context of business service schema should expose it lineage to the master schema/data model (or at least identify the version of that data item). Further that each item has a clear and explicit semantic that isn't changed in any way through it use in the specific context of the service's message (that is, a customer as defined by the core business domain model is still a customer in a message, an address is an address yadda yadda). However as you say, the question of namespace visibility then arises as does the 'total' global element/type (aka. garden of eden) approach to schema design versus other models which enable reuse but allow namespace 'hiding'. Some seem to prefer the elementForm='unqualified' ? Fraser. On 08/09/06, Michael Kay <mike@s...> wrote: > > > > So are you talking about auto generating a bunch of schemas > > based on a core description of the data model and a set of > > rules for getting from the core description to each > > individual case you need? Or am I misunderstanding? > > Yes. But perhaps not so much a core description of the data model, more a > set of reusable and parameterizable components that can be included in > messages, combined with a description of each message in terms of which > components to include (as required, optional, or repeating). > > One area where I'm still very ambivalent is whether different messages > should use the same namespace, given that they might apply different rules > to the same elements. > > Michael Kay > http://www.saxonica.com/ > >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] |
