[dsdl-discuss] Re: Semantic framework for namespace processing and MNS2

> I like this idea. If I implement DSDL-VC, I would probably create
> a SAX filter for implementing this data model.
>
> The last paragraph in this section puzzles me. Consider a document
> <foo xmlns=""><bar/></foo>. How many element sections does this document
> have? One or two?

I can see why it's puzzling. In the data model, an element section
contains its child sections and an element contains its child elements. So

<foo xmlns=""><bar/></foo>

is one element that contains one child element. It corresponds to one
child element section that contains no child element sections. I can
construct an element section corresponding to <bar/> even though this is
not required for constructing the element section for <foo
xmlns=""><bar/></foo>.

> Your approach is algebraic or functional. However, I am not sure if
> such an approach facilitates stream-based implementations. I think
> that this functional or algebraic approach suggests DOM-based implementations
> too strongly.

Stream-based implementation is an important (I would say essential)
requirement. However, I don't see that the kind of formalism used to
specify the semantics has much impact on the possibility of stream-based
implementation. What's important is the nature of the semantics
specified, regardless of what formalism is used. I've just completed a
stream-based (SAX) implementation of the semantics that I've described,
so I don't think there's any problem for my proposal in this area.

> I have been thinking about a logical formalism. Now that I have seen your
> data model, I think that it is possible to create a logic-based formalism for
> studying VCSL, Namespace Switchboard and MNS. I am now writing a note about
> this and will send it to this mailing list tomorrow.

I think the best would be to describe the same semantics using multiple
formalisms. Different formalisms often provide different insights into
the semantics.

James

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)