Comments on R2RML draft

I spent some time with the draft over the weekend.  Comments below.
Please do not close Action-83.  as this will remind me that the draft needs to be reviewed
more thoroughly and periodically as it evolves.

MAJOR

- The subject in the RDF triples comes from the primary key of the logical table.
We don't say that.  Also, the URI for the subject is (usually) constructed from the value
of the primary key but may be constructed by other means.  We also don't say that.

- Should we say something about using "well--known"  URIs e.g. from OKKAM.  See usecase
from Angela.

- What happens if the table does not have a primary key?

- We need more words around the examples to explain what is going on.  For example:  This
SQL query selects ... and constructs a URI ...

MINOR

- The status section and the intro refer to the Direct Mapping document in the future tense .  This needs
to be changed

- Can we show a simple example - perhaps mapping from the EMP table -- early in the document
to give people a flavor of what the language looks like?

- There are issues in the document that are not in the issues list.  Need to fix that,

- We need to say somewhere, early, that R2RML is only for query not for update.

- Should we define TriplesMap before TermMap?

- The first line of A2 uses the term "the RDF".  There is also a typo "specificied".
-- 
All the best, Ashok

Received on Monday, 31 January 2011 13:59:08 UTC