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


Help: OASIS Mailing Lists Help | MarkMail Help



   Inferencing across XML types was Re: [xml-dev] XQuery types

[ Lists Home | Date Index | Thread Index ]

Dare Obasanjo wrote:
> From: Jonathan Borden [mailto:jborden@attbi.com]
> Sure. What I am saying is that this "freshman philosophy
> course" way of looking at classes _actually has_ real world
> utility (and is implemented in real world production software btw).
> For example, I would like to state:
> widgitCo:purchaseOrder  --daml:sameClassAs-->  my:purchaseOrder .
> i.e. these two purchase orders are the same.
> Now they might have different namespaces, different element
> names etc, but if the computer could understand that they are
> the same (much as a human clerk could easily see), this would
> be a really powerful "use case" for semantic
> interoperability. Done correctly this sort of thing is not
> some pipe dream rather within our reach.
> That point is far away from XUpdate, but the type of thing
> (sic) that I'd like to see the community work toward.

I see where you are going with this but I'm not sure the type system
described in the XQuery formal semantics will get you there.

On the one hand, you can already do this in much the same way that the
XQuery formal semantics does. Basically widgitCo:purchaseOrder is the
same type as my:purchaseOrder if they have the same W3C XML Schema type.

On the other hand, if you are looking for inference rules to help
determine this without the use of W3C XML Schema then I guess this is
what the semantic web folks are working on which I've derided on
multiple occassions as being a pipe dream akin to the "XML will make
search engines smarter" hype of yesteryears.


No direct need for XML Schema at all (except for simple datatypes)

suppose the two XQuery (formal) types:

type widgetCo:purchaseOrder {
        element name {xsd:string},
        element price { xsd:int }


type my:purchaseOrder {
        element myName {xsd:string },
        element myPrice {xsd:int }

and a few statements (in N-triples form i.e. subject predicate object):

widgetCo:purchaseOrder daml:sameClassAs my:purchaseOrder
widgetCo:name daml:samePropertyAs my:myName
widgetCo:price daml:samePropertyAs my:myPrice

then yes, an engine can inference that widgetCo:purchaseOrder is the _same
as_ my:purchaseOrder even though they don;t have the same XML Schema type.

No pipe dream.

In any case, there is alot of loosey goosey mumbo jumbo pie in the sky stuff
written about the so-called "Semantic Web". On the other hand if we define
the term _Semantic_ properly, then yes there is actual code that can make
hopefully useful inferences - today.



News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS