[
Lists Home |
Date Index |
Thread Index
]
<datahead>
There seems to be missing any notion of the abstract information model from this
discussion.
For any application, there is a hypothetical domain model, which is a
representation of all knowledge of that domain. It is unlikely to exist in any
complete realised form, but only in the heads of the experts in that domain.
From this is created and abstract information model, which has the information
types, structures, relations and mappings for some (hopefully useful) subset of the
domain.
Any application within a domain is a mapping of an AIM to the concrete processes
and structures of an application language.
For XML based applications, there is a mapping of the abstract data into a concrete
syntax defined by the XML spec and constraints in a schema language.
Part of this syntax allows the elements in this mapping to be grouped into a
namespace.
To me, the best use of the namespace is to identify a particular mapping of the AIM
to the syntax, not a mapping of a collection of syntax constraints to a document;
the namespace of an element identifies the AIM of the element.
If I have an element from the UML/XMI namespace then I can create an application
which will assume that there is some data stored a certain syntax which maps onto
information in the UM-language that can be manipulated in various ways which appear
meaningful to users conversant with that language. This is invariant against
whether the element is embedded in another document, or what that document's schema
is (as long as the document is valid against its schema and the schema's definition
of the element's structure is allowed under the XMI mapping); though such
contextual information is useful in other terms (for example, on one system I have
designed, an UML fragment embedded in a change note is used quite differently to
one embedded in a pattern template, but the fragment's behaviour and meaning are
the same).
What would you want from a RDDL directory found by resolving the UML/XMI namespace
URN?
Links to every document type that can embed UML/XMI?- impossible
Links to every tool that can process UML/XMI?- difficult
Links to one tool that can process UML/XMI?- political
Links to syntax description languages for UML/XMI?- possible, but the XMI doesn't
have unique mappings from information to data, so how useful?
It is _very_ hard to define, in a machine readable manner, that a syntactic element
maps correctly onto a domain knowledge entity.
It is possible to define, in a machine readable manner, that a syntactic element
encodes an abstract information model entity.
If my XML namespace points to something which defines, say, how it encodes a subset
of the ISO-STEP component assemblies abstract information model, then that is
potentially _very useful_ and means that any application which is based on that AIM
and can parse generic XML into its structures (given the mapping) can do useful
things with it.
(If there are not international standards for AIMs for invoices, then that's a
different problem ;-)
Pete Kirkham
(engineering [meta]* modelling pedant)
</datahead>
********************************************************************
This email and any attachments are confidential to the intended
recipient and may also be privileged. If you are not the intended
recipient please delete it from your system and notify the sender.
You should not copy it or use it for any purpose nor disclose or
distribute its contents to any other person.
********************************************************************
|
