- From: Patrick Stickler <patrick.stickler@nokia.com>
- Date: Fri, 4 Oct 2002 10:54:21 +0300
- To: "RDF Core WG 7332#" <w3c-rdfcore-wg@w3.org>, Patrick Stickler <patrick.stickler@nokia.com>
Rather than seeing growing consensus in the past week for either of the two proposals on the table, we now have two new proposals on the table and a plethora of questions as to how they compare to the earlier proposals and to the requirements. I would like to suggest again that it does not appear likely to me that this WG will achieve any form of strong concensus on this issue, and that if a decision is forced, it will be one of weak concensus, severe compromise, and for many made out of exhaustion induced apathy. That is not a good context for deciding on an issue which will have significant impact on large numbers of present RDF users and impact all future RDF based models and systems. We have laid the conceptual groundwork for RDF datatyping and provided a mechanism for explicitly datatyped literals. Of that the WG has had nearly unanimous agreement. We should defer the issue of inlined literals to a future WG. This means that no existing applications must change, and yet interoperability can still be achieved for the time being by exporting graphs with all application specific datatyping assumptions expressed explicitly by datatyped literals when sharing knowledge with other systems. No, this is not the ideal final solution, bit it will work for now and does not burn any bridges or close any doors to future research and implementation experience that may be brought to bear on future attempts to provide an answer to the question of what inlined literals should mean in the RDF MT. I ask again that this proposal be considered by the WG during the telecon. Patrick
Received on Friday, 4 October 2002 03:56:07 UTC