Re: In RDF what is the best practice to represent data provenance (source)?

Hello,

I've been reading for a bit and am new to actually commenting.  I enjoyed the conversation on reification and would like to add algorithms to the list of things to consider in provenance tracking.

We might imagine a semantic web service that allows users to post SemanticWebPaper objects and imagine that this service returns an error to the sender in the event that the logical conditions aren't met defining the class.  Will this service trust the incoming document's list of keywords or perform its own processing on the document to generate a keyword list for checking this logical constraint?  Assuming the service reprocessed the keywords from the document, it should conceivably ammend the OWL object to differentiate the OWL from the sender and the OWL from the algorithm when it puts up the document in a user environment.

In my imagination, at least, its entirely possible that structured documents will be modified by numerous tools we make as they move around the Internet.  So, I'm stopping short of describing a 'wiki-graph' scenario but I do feel provenance tracking is an important issue.

I have a question regarding the human-writeability factor.  Let's say I have a candidate syntactic solution for reification of RDF-based documents that allows the document to remain one graph for processing and allows some statements to reference statements or subgraphs in the same document, at the same time, but the downside is that notepad or vi wouldn't work for composition anymore-- some simple software for editing would be required.  Would this idea meet a lot of resistance?  I'm new to the discussion and don't have a handle on how important human-readability, human-writability is to the community in weighing all these considerations.


Cheers,
Adam