Comment on Format Proposal

[Olaf Kummer <kummer@informatik.uni-hamburg.de>]

Dear Michael, dear all,

I'd like to comment on your XML format as proposed in 
the forthcoming paper for the AWPN available at
  http://www2.informatik.hu-berlin.de/top/pnml/download/JKW_PNML.ps

The terminology section is nice and clear. (I'll try to stick to
it, I really try.)

I can accept the page/reference architecture as it is described in
the paper. (I am not commenting on the detailed XML Schema at
  http://www2.informatik.hu-berlin.de/top/pnml/detail.html
because it is dated 2000/06/26, so it might not represent your
views accurately.)
 
The representation of graphics and the basic structure of nodes
and arcs seems to be common sense. Giving the arc type 
as an attribute in a separate element might be just as good 
as other alternatives, hence I'd like to leave this as it is now.

I still have one problem with the annotations. Their type is
hidden in the name of the elements. That looks very general,
but it can be a burden for graphics-only tools. These have no
way of determining the graphical display of an annotation
without having detailed knowledge of the net formalism.

E.g., a tool sees
<place>
  ...
  <foo>...</foo>
</place>
Is foo an annotation of type "foo" or an extension of the XML
format, e.g. an analysis result regarding this place? A tool
could not know without looking at the conventions document,
which might be unavailable.

I know that my initial suggestion to write
<annotation type="foo">
  ...
</annotation>
was not received happily, probably because it does not allow
to specify the content of the annotation element precisely.
(This applies to both DTDs and Schemas.)

Hence let me suggest another mechanism. Write:
<place>
  ...
  <placeAnnotations>
  <foo>...</foo>
  </placeAnnotations>
</place>
Now it is clear to all tools that every element contained
in the placeAnnotations element will be an annotation. We can also
aggree that such an annotation has the general structure proposed 
in the paper:
<place>
  ...
  <placeAnnotations>
  <foo>
    <value>...</value>
    <graphics>...</graphics>
  </foo>
  </placeAnnotations>
</place>
Using this architecture, the graphics tools can work independently
of the net formalism. Would that pose any problems?

In general, I like the proposal a lot and it looks very
promissing.

Regards, Olaf
-- 
Olaf Kummer, Luruper Weg 21, 25469 Halstenbek, Germany
Tel: 04101-473957 / 040-42883-2245    Fax: 040-42883-2246
mailto:kummer@informatik.uni-hamburg.de
http://www.informatik.uni-hamburg.de/TGI/mitarbeiter/wimis/kummer.html