Re: The less abstract Infoset

[Jeff Rafter <jeffrafter@earthlink.net>]

Jonathan,

Thanks for the reply.  Suddenly my question seems rather naive.  I am
wondering though, are these implementations of the Infoset or do the simply
make use of the Infoset?  It seems that an "implementation" of the Infoset
might define a simple class for an information item which essentially wraps
a Hash table of properties and values.

SAX could emit objects such as these when parsing
DOM could use them to build the tree
XPath could walk through them
Schemas contribute to them.

I know I am basically explaining how everything already works here (to the
man who already knows no-less!) But I thought that was the push for all of
the Version 2 specs.  The Infoset would be the spoke between everything else
and the custom data model expressed in all of the above specs would be
abandoned (or at least conformed).  One could use these little data packets
(e.g. classes) to allow implementations of different specs. to communicate
in one standard way.  My question was intended to find out if anyone has
implemented the information items from the Infoset verbatim (like in an
InformationItem Class) that would be the spoke and if so how?

I know that it is defined as an abstraction and that the Infoset takes no
stance but I thought maybe somebody took this route-- I also know I might
also be digging myself deeper here.  Either way, I know I am missing
something potentially obvious.

Thanks,
Jeff Rafter


----- Original Message -----
From: "Jonathan Borden" <jborden@mediaone.net>
To: "Jeff Rafter" <jeffrafter@earthlink.net>; <xml-dev@lists.xml.org>
Sent: Friday, March 09, 2001 3:53 PM
Subject: RE: The less abstract Infoset


>
> Jeff Rafter wrote:
> >
> >
> > I have been looking at implementing the Infoset recently but have found
> > that, as I move deeper into the 'abstraction',  I seem to
> > gravitate less to
> > the 'abtract' and more to the 'vague' nature of the Infoset.  I am
looking
> > for hints/tips on real world implementation of the Infoset in the way of
> > interfaces or strategies for representing the information items.
> > Some of my
> > biggest confusion centers around speed/memory and extensibility.  I
> > apologize in advance if this question is vague, asking for too much...
> >
>
> 'real world' implementations of the Infoset roughly include:
>
> SAX
> DOM
> JDOM
> XPath
>
> The infoset is intended only as a model or abstract syntax to which an API
> conforms.
>
> -Jonathan
>