[XML-SIG] I am confused!.. can I pick someones brains on parsers?

[Ken MacLeod]

"gary cor" <stuff4gary@hotmail.com> writes:

> It seems to me that there are loads of parsers in python for XML
> that do the similar things, I wonder if someone could try and
> explain to me how parsers differ from Dynamic Linking Extensions on
> a PC or a Macintosh extension files and developing for these
> systems.  Is it not possible to have one parser for image files
> (like a image gear DLL), another for dealing with text data, another
> for XHTML, another for SVG, another for movies, etc...  My point
> being can't parsers be more application specific?

Yes they could, and some (grovesy folks*) would argue that there
should be lots of them.  That's not the current, common practice
though.

> And if they are application specific how do I find which ones I
> should use where and the way each is to be used for its task within
> its application?

Like the mapping from mime-types to applications in browsers, there
would also be some kind of configuration that would map mime-types to
their respective parsers.  All this is just brainstorming, though,
since no such system exists as yet.

There recently has been a new specification out called Resource
Directory Description Language (RDDL) <http://RDDL.org/> that may be a
great help in mapping data types to resources that apply to those data
types (parsers, style sheets, documentation, validators, services,
etc.).  It's currently focused on providing resources for XML
Namespaces, but the parallel to any type of data is very clear.

  -- Ken

* The term "groves" refers to an abstract way of modeling data as
nodes and properties (using objects and attributes in Py), with the
ability to do processing on any set of nodes in a common way.
Resource Description Format (RDF) is a similar abstract model with
roughly the same characteristics.