Re: Review of RDFTM Survey

Hi all,

This message provides my comments on the editor's draft of the Note 
"RDFTM: Survey of Interoperability Proposals".

I have read Natasha's comments, but reviewed the document beforehand.  
I apologise for any duplication.

I am not an expert in Topic Maps (and not really in RDF, either!).  
Please take my comments accordingly.  It is important, though, that the 
document be accessible to both communities and not just to those with a 
complete understanding of both.

OVERVIEW
--------

Overall, I really liked what this document provides.  I think Natasha 
is right that a proper review of the state of play would be 
academically publishable with the right preparation.

The document seems to focus on comparisons between RDF (and 
occasionally RDFS) to TMs.  Has any work been done to add OWL to the 
comparisons?  Would it be useful to do so?  My rather naive thinking 
suggests that it may be.

Some of the referenced authors note that RDF and TM may not be directly 
mapped.  Do the editors agree or disagree with that?  After reading, I 
was left feeling that they had chosen not to say, while being 
optimistic.  Is that the case?

DETAILS
-------

The title really needs to change.  "RDFTM" is an acronym used only 
internally by the SWBPD WG.  I would suggest something like, "A Survey 
of RDF/Topic Map Interoperability Proposals".

The W3C itself did not initiate the RDFTM task force, the SWPBD WG did. 
  The WG should be named in the Status and Background sections.  This is 
especially important since any Note coming from this work would be an 
SWBP document :)  Similarly, the RDFTM TF itself should not be 
referenced in the document, since it is a transient part of the 
publishing WG.

I don't think that WG *intentions* should be put into Notes.  
Therefore, this document should not name future deliverables or refer 
to them.  Ralph may suggest differently?

There are some terms that are used and not defined first (such as 
RDFTM, RDF2TM, TM2RDF).  They may be obvious to us, but we shouldn't 
assume that the average reader will find them understandable.  Are the 
terms TM and XTM synonymous?  They appear to be used that way, but I am 
not sure.

Under the Unibo description, the document says, "implemented in Meta".  
What is Meta?  Is there a reference to it?

One should be very cautious when naming commercial products, even if 
they are relevant.  I suggest that commercial products be named solely 
in references.  We need to ensure that we do not appear to be endorsing 
any particular product, EVEN IF IT IS THE BEST THING OUT THERE.

There are several places where the document says that a particular 
feature is "not possible in RDF".  I was confused by this in some 
places.  RDF, RDFS or RDF/RDFS/OWL?  Surely multiple options for a 
given label may be represented in RDF; what are the ramifications in 
relation to RDF/TM integration?  I was left without an answer since the 
dismissal of the issue occurred to early.

The Test Cases for each proposal are dense and take a lot of space.  
They are probably not (completely) accessible to those without a rather 
complete background in both RDF and TM.  Perhaps these can be split out 
into separately-published URLs?  Alternately, they could go into an 
appendix.

It was not clear how the test results related to one another.  I had to 
flip around a lot to get that overview.  A summary table would help 
readability.

The conclusion is reasonable and has the right content, but it might be 
nice to expound just a little bit (in spite of the length of the 
document).


Good job to the RDFTM TF!

Regards,
Dave

Received on Wednesday, 2 March 2005 22:26:50 UTC