Re: RDFa and RDF API, or Semweb and JavaScript

Hi Niklas,

Niklas Lindström wrote:
> On Thu, Apr 28, 2011 at 2:42 PM, Nathan <nathan@webr3.org> wrote:
>> Thomas Steiner wrote:
>>>> I recently read the RDFa API [1] and RDF API [2] and quite liked them
>>>> (nice work there, seriously).
>>>> I wanted to ask if there are already decent implementations of those in
>>>> JavaScript?
>>> Somewhat dated, but the best implementation out there so far:
>>> https://github.com/webr3/rdfa-api.
>> Cheers Thomas, aye that one is from an previous version where the RDFa API
>> and RDF API were defined as one API in one doc; have a new RDF API specific
>> one pretty much ready to go in the next couple of days.
> 
> Great; I'm looking forward to it!

Likewise, enjoying getting back in to productive coding at the minute - 
fingers crossed it should be finished this evening :) HOWEVER! (that's a 
big however)...

>> Thinking about pulling the js3 [1] native handling functionality in to the
>> implementation.. anybody think that's a good idea? (it can be done whilst
>> still maintaining full API compatibility)
>>
>> [1] https://github.com/webr3/js3

On thursday we (the RDF Web Applications WG) decided to create two 
documents, the former RDF-API is now moving to a document called the 
"RDF Interfaces" and the RDF-API is going to be a second level API, 
similar to the RDFa API, focussed at end developers, somewhat of a 
simpler read only API, including things like Project and query mechanisms..

As per your note below Niklas (or anybody else interested in the space) 
I'd encourage you to give Manu Sporny, Ivan Herman, Benjamin Adrian or 
Myself (or the list, which is still the rdfa list) a shout if you're 
interested in joining in and helping the process along, I'm sure the 
input would be very much appreciated, invaluable even.

> IMHO, I think it could be wise to leave it out. I'd prefer a base
> package with code (and ideally tests) implementing the spec and
> nothing but the spec. While JS3 is very nice, I expect that different
> solutions to those needs might crop up, and I find it sensible to have
> it as a separate package upon the core API(s).
> 
> That way, it's also reasonable to bless your implementation as the
> reference implementation of the spec, and make it clear for authors of
> ports to different languages/platforms that this code covers nothing
> beyond the standard API. Also, it allows for JS3 to develop and take
> shape independently.

Good point, the only minor detail is that JS3 features using ecmascript 
native types as the core RDF terms (typed literal, plain literal etc) 
and thus it's more a layer under than above - however, I agree fully, 
best to keep the JS3 functionality in a different lib.

Best,

Nathan

> (This is the way e.g. the RDFLib community has worked, with a core RDF
> API (and parsers), upon which different "O/R" mappers like Sparta,
> Oort, RdfAlchemy and SuRF are built, each complete with their/our own
> needs, ideas and idiosyncrasies.)
> 
> (Also, as I said elsewhere, I'd be happy to pitch in and implement
> e.g. Projection and some tests/specs if you're short on time. ;) )
> 
> Best regards,
> Niklas
> 
>

Received on Saturday, 30 April 2011 11:50:43 UTC