RE: Proposed new structure for SIMILE [Relevant] Technologies doc umen t

Hi Rob
 
> Some open issues:
> 
> * Document title:  Some of the document describes problem 
> areas, some describe technologies.  How about "SIMILE 
> Problems and Relevant Technologies"?

This seems to describe the document better, but its still a bit of a
mouthful so I suggest we adopt this *unless* someone has a better proposal?

> * Although I like Andy's 4 layer structure (SWeb platform 
> player, Content Layer, Domain Layer, Client Layer) it doesn't 
> strike me as being the best structure for this document.  
> Perhaps at some point we could add a section describing this 
> model and relate to it in other parts of the document, though 
> I don't know if we'll have time to do this before the plenary.

My preference would be for a section discussing the 4 layer structure rather
the structuring the document to reflect it. 

> * There are a few loose ends I'm not sure where to fit in.  
> "Classification" (currently section 3.10) seems to be talking 
> about both simple typing in the RDF sense and classification 
> in the (semi-)automated categorisation sense.  

Yes, this is something we argued about at the meeting where we were meant to
travel but in the end ran stuff by videophone? So the document reflects the
fact that the team didn't agree and that we wanted to ascribe two meanings
to that word - see
http://lists.w3.org/Archives/Public/www-rdf-dspace/2003Apr/att-0061/pi-meeti
ng-2003-04-09.txt

> Also 
> "Necessary and sufficient constructs" (currently 5.2), 
> and 
> current section 6.7 "Misc" seems to have all sorts of odds 
> and ends.  

Perhaps these sections should be moved into the Introduction document, into
either the opportunities or challenges sections?

> For now I've put in a catch-all 'Miscellaneous 
> other issues' section, if someone has a better idea let me know.

We either need to flesh out some of these issues, or get rid of them. They
stem from the mindmap, but I wasn't at all the knowledge capture sessions
that generated the mindmap so I don't know how some of them should be
fleshed out. Unfortunately I think we need Mick or the people responsible
for those points to flesh them out. 

> * In the Metadata section below, the line between 'Creation' 
> and 'Augmentation' is somewhat blurred -- augmentation could 
> be considered a special case of creation.  However I added a 
> separate Augmentation subsection.  Can anyone offer a better 
> structure?

I didn't like augmentation, as I'm against creating terms. However I did ask
people to explain it and the explanation I got was there are three
situations (this may be slightly different to what I said before so
apologies in advance)

- the metadata comes from the content (extraction)

- it comes from some kind of process e.g. rules or a machine learning
classifier (augmentation)

- it comes from a user (creation)

However there was some discussion on this - see
http://web.mit.edu/simile/www/documents/researchDrivers/rd_issues.html#b1b1c
52

> Thanks all.

> Processing Models
> 
>   (as current, Mark B working on some edits)

Do you need me to work on edits beyond the ones I proposed?

Dr Mark H. Butler
Research Scientist                HP Labs Bristol
mark-h_butler@hp.com
Internet: http://www-uk.hpl.hp.com/people/marbut/

Received on Tuesday, 1 July 2003 10:55:04 UTC