Re: Proposal for the datatype mapping

Richard - Thanks for writing this up, this looks promising to me. A couple
of comments inline below.

On Tue, Nov 22, 2011 at 8:16 AM, Richard Cyganiak <richard@cyganiak.de>wrote:

> 1. The spec formally defines two kinds of RDF literals corresponding to a
> SQL value: the “natural RDF literal” and the “canonical natural RDF
> literal”.


I like the idea, "canonical natural" strikes me as confusing words. I will
ponder what I might propose as an alternate label for this.


> 4. When generating IRIs in R2RML, the output SHOULD be the canonical
> natural RDF literal
>

I take this to mean: when generating IRIs from SQL values, the string
representation used for the SQL values in the IRI SHOULD conform to the
lexical value of the "canonical natural RDF literal" for the SQL value. Is
that what you mean?


> 5. It is noted that R2RML mapping authors SHOULD explicitly convert
> non-string columns to string in an R2RML view when generating IRIs to
> maximize portability of their mappings.
>

And if the mapping author does not explicitly convert non-string columns to
string, the R2RML processor will implicitly convert them, and perhaps not
use the "canonical natural RDF literal" form and therefore different R2RML
processors may produce different IRIs for the same mapping and data. Just
re-iterating what I think you are saying to make sure I understand the
proposal.