Serendipitous Interoperability

I wrote the following to Jim Hendler and Ora Lassila.

Jim requested that I post it here.

Thank you for your attention,

Jim Farrugia


---------- Forwarded message ----------
Date: Fri, 14 Dec 2001 10:42:56 -0500 (EST)
From: Jim Farrugia <jim@spatial.maine.edu>
To: Ora Lassila <daml@lassila.org>, hendler@cs.umd.edu
Cc: Jim Farrugia <jim@spatial.maine.edu>
Subject: Re: Serendipitous Interoperability

Ora and Jim,

I just read Ora's posting to webont
http://lists.w3.org/Archives/Public/www-webont-wg/2001Dec/0121.html
and I have some comments.

If you think I should post these to the rdf-logic list, let me know.

I think Ora hit the nail on the head when he said many simple examples
can be solved by other (non-Semantic-Web) means. If the WOL is to take
hold, I think as much effort needs to go into documenting the "why" of the
language as goes into developing the language itself. Specifically,

 * several uses cases that demonstrate in detail why a WOL is needed
   to solve them.  These use cases may be different from the use cases 
   being presented now, or there may be some overlap. But the difference 
   in _intent_ would be that while the use cases now are presented as "here 
   is what I would like a WOL to allow me to be able to do," the use cases 
   I am suggesting would say something more like "here is an example of 
   why a WOL is needed to solve a particular problem."  Also, such use cases 
   would be small packaged examples written to an audience of language 
   users, not language developers. Specifically, the following structure 
   seems useful to me. 

   Start with a description of a general use case.  Give several specific 
   issues/problems that this use case treats.  Explain which of these 
   specific problems can be solved w/o the WOL and why/how.  Explain which 
   of the specific problems can't be addressed without a WOL and why 
   not.  I think this latter piece is the most critical part.  Then show 
   how certain constructs or features of the WOL solves those specific 
   problems.  I think that all this could be written up in 2-5 pages.  
   And I think the webont group would want to have, say, at least a dozen 
   such cogently argued cases.
   
   Do you think any part of the webont group should/will take on these
   responsibilities? It seems that the language developers must have
   already internalized many of the reasons why a WOL is  needed. If 
   they, or another subgroup (publicity???:-) could be charged with writing 
   short pieces specifically addressed to potential users of the 
   language, it seems it would help with the adoption of the WOL.  

   I think what I am suggesting represents a concrete way of addressing
   the problem that Oro mentioned (about simple examples ...).

   What do you think?


I also have a set of comments about how the webont group suggests
that the language might be used - in the sense of how will 
documents/resources get tagged with WOL?  What kinds of tools are needed,
and how are they different from existing editors?  How can semantic
issues of WOL be addressed in an editor?  I'll leave my detailed comments
for a subsequent email.

Thanks,

Jim

Received on Friday, 14 December 2001 10:55:06 UTC