An example reification of AIF
Matt South
on in various places behind the scenes. One of the things that we've
been doing at CRUK/ASPIC is an XML reification of AIF that I think is
now ready to share. That's not to say that it's finished or anything
- there's still a long way to go, but I think there may enough there
that's of interest to other people.
Details of the XML Schema's that we've been using in ASPIC, based on
the AIF strawman, can be found at http://aspic.acl.icnet.uk. I had
hoped to publish some nice script that consumed the AIFXML we generate
and creates some nice graphs (a top-level argument network and a
complete argument detail + network are the first candidates) but
unfortunately my other priorities havent allowed this as yet. In ASPIC
we've used the extended schema for exchanging proofs between the
Inference software and the Learning software. The next steps for us are
to demonstrate it's use in the context of decision making followed by
dialogue. Note that these schemas don't yet take into account the
recent discussions (in regards to the KER paper) on the differences
between attack and preference.
regards,
Matt
as a taster - here's some of the (top level) AIF.xsd documentation
annotation...
This schema is a high-level, first-cut implementation of an xml-schema
that's within the spirit of the the Argumentation Interchange format as
described in the Strawman document (1). AIF is intended to facilitate
the transmission of arguments between different end-points, including
software agents and argumentation visualization tools.
My original goal in this exercise was to define,
* a general schema (loosely describing a basic structure for an XML
document) and
* a particular schema that overides the general the schema for a
particular implementation.
Thus if I created a particular application that generates AIF then I
could demonstrate a schema for that implementation building on some
basic principles and any aif generated by that implementation should
validate against both schemas.
However, upon closer examination it turns out that the only way to do
this is in XML Schema is to add in "extension" nodes into the schema
which can then allow multiple "xs:any, namespace:other" children (2).
This will work, but to my eye, is not at all elegant, especially after
more than one layer of inheritance. So the extensibility mechanism in
this design is xs:extension and an example of an extended schema is
shown in AS.AIF.xsd that is included in this directory. An extended
schema should first include this schema and then use xs:extension to
add further attributes and children to context, s-nodes, i-nodes and
maybe edges.