Substantial comments on the DM Mapping

Eric, Juan,

as promised, these are the more substantial comments. I am no 100% how you two divide up the work; I would expect most of the comments would be handled by Eric except for the very last one...



- The status of the document should reflect that this is last call. Something like that should appear in the status section (text stolen from another document):

This is a Last Call Working Draft and thus the Working Group has determined that this document has satisfied the relevant technical requirements and is sufficiently stable to advance through the Technical Recommendation process.

- It would be good to share the characterization of the direct mapping and r2rml and how these two compare (answering the almost inevitable question that the community would/will ask: why these two?). 

The R2RM document currently says:

This specification has a companion that defines a direct mapping from relational databases to RDF [DM]. In the direct mapping of a database, the structure of the resulting RDF graph directly reflects the structure of the database, the target RDF vocabulary directly reflects the names of database schema elements, and neither structure nor target vocabulary can be changed. With R2RML on the other hand, a mapping author can define highly customized views over the relational data.

The same statement (well, modified for the DM case in terms of 'companion') should maybe added to the DM document. The text that is currently there is really not enough.

- Section 2. The text currently says:

"...RDF graph that is called the direct graph. This graph is composed of relative IRIs that may be resolved against a base IRI"

which seems to suggest that the RDF graph will have relative IRI-s in their nodes. I do not think that is correct: an RDF graph always uses absolute IRI-s. The usage of the base IRI is part of the generation process and/or the specific serialization of the graph, and not part of the final graph...

- Section 2.5, generated turtle example: with the usage of @prefix, the last few statements yield the subject as:


Note the '_' character at the end. This contradicts 2.3 which would yield:


I think the one in 2.5 is wrong.

(b.t.w., as this is an informal section, it may be worth noting that the '+' character in the URI stands for the encoding of the space character)

- Section 2.5: I do not think it is appropriate to keep issues in a Last Call WD. The second issue is actually moot now, with both formalisms in the document. The first issue should also be removed.

- Section 3, definition row node: it says 'the row node is a relative IRI...'. This is the same remark as above for the graph: a node in an RDF Graph has an absolute IRI; the usage of the relative URI is a generation/serialization artefact. What about

"the relative IRI to define the node..."

or something like that.

- Section 3, the definition of the property IRI seems to contradict the examples. Taking the very first example, the property IRI-s in the example are of the form

... <http://foo.example/DB/People#fname> ...

whereas, if I follow the specification here in section 3, I think I would get:

... <http://foo.example/DB/People/fname#> ...

I think the right one is the example, which also coincides with entry [48] in A.4...

Also, I am not sure the definition should refer to a primary key. The property IRI generation should be valid for any column, and the definition reads as if only primary keys were used...

- Appendix B: the cross links should be redone. The target for the link to Column IRI has been changed, the same for Table IRI, etc. Essentially, all cross-links should be checked, and should point at the normative definition.

Ivan Herman, W3C Semantic Web Activity Lead
mobile: +31-641044153
PGP Key:

Received on Monday, 8 August 2011 11:19:48 UTC