- From: Tim Berners-Lee <timbl@w3.org>
- Date: Sat, 23 May 2009 16:12:37 +0100
- To: Dan Brickley <danbri@danbri.org>
- Cc: Renato Iannella <renato@nicta.com.au>, Harry Halpin <hhalpin@ibiblio.org>, "martin.hepp@ebusiness-unibw.org" <martin.hepp@ebusiness-unibw.org>, Peter Mika <pmika@yahoo-inc.com>, Sandro Hawke <sandro@w3.org>, Ralph Swick <swick@w3.org>, "www-archive@w3.org" <www-archive@w3.org>, Libby Miller <libby@nicecupoftea.org>, libby.miller@bbc.co.uk
Dan, You make the point strongly that a few simple terms in FOAF have a great use, and there is a limited damage to the world if they overlap some more complex and complete ontology in the same space. However, in this case I don't think we have a need for something much simpler than VCard. There is a level of coverage necessary for an address book, which is the target of many efforts, to name a few :- - The 2000 contact ontology I made and used as a lingua franca http://www.w3.org/2000/10/swap/pim/contact - The old VCard ontology -- consensus seems to be this is obsolete. (uri? -sorry on a plane) - The 2006 VCard ontology http://www.w3.org/2006/vcard/ns - The 2007 LDAP -- derived ontology equivalent to the LDIF format http://www.w3.org/2007/ont/ldif# The LDIF format is exported by Thunderbird Addressbook for example. See converter http://www.w3.org/2000/10/swap/pim/ldif2n3.py and my 2007 notes eg http://www.w3.org/2002/12/cal/vcard-notes.html - Whois++ derived ontology? - etc etc .. hcard ... etc It seems to me the scope is pretty well defined, except for a few areas where we will need to trade off being simple vs being complete, and being clean ontology vs being able to round trip with vcard. This tradeoff will I think have to be done term by term. So what I am saying is that this is a very clearly needed ontology, but the people who use it won't all want to use the rest of FOAF. The number of people who need to be involved in designing it are larger than those defining some other parts of FOAF. I think the FOAF process, of Dan and Libby being benevolent attentive authors fo an evolving spec, worked because for social networking especially foaf:knows your were way ahead of the pack. I don't think the same process will work in this very populated area. We will need experts who have understood each of the ontologies (vCard primarily and secondarily LDIF and teritiarily whois++). We will need test suites and conversion programs and test harnesses and serious error tracking. I think the idea of having yet another simple ontology within FOAF will add to the confusion and the already rife fragmentation. In fact I sympathize, danbri, with your goal of having a basic profile of a person in just the FOAF namespace. I think the only way you will get it is the FOAF community work in a more formal process to make a solid fairly complete ontology for those terms which relate to Vcard. Now, FOAF has pioneered the idea of different terms in an ontology being in different levels of stability. So one could imagine different terms being in different processes, different groups. Would that work? Tim
Received on Saturday, 23 May 2009 18:33:30 UTC