Re: RDF API convergence? was Re: ANNOUNCE: RDF.NET

>>>"Alberto Reggiori (vaio)" said:

> Personally I have been looking for such an RDF API spec when I
> started off with RDFStore [1], and at that time I ended up using
> the great work of Sergey Melnik.  Today we have a larger and richer
> set of RDF API implementations (and understanding), and I also
> personally feel the need to converge, and sit down and try to
> specify such a document.
> 
> A list of requirements could be the following:
> 
> - specified using IDL interfaces

Before starting Redland's scripting interfaces I looked around and
found IDL sufficient for describing pure OO APIs but ultimately not
as useful as SWIG http://www.swig.org/ which allows you to define the
language independent interface and generate many scripting
interfaces.  If we pretend C is the portable assembler language of
the net.

SWIG allowed me to mechanically generated interfaces in Python and
Perl and just works.  A really useful tool.  In the long term I could
see generating such interfaces from *an* IDL, not necessarily IDL, if
that makes sense.

> - bindings into the major languages (Java, Phyton, Perl, C, C++, C#
>  and so on  )
> - layered - e.g. core, parsing, storing, querying, services and protocols
> - support either statement and resource centric views
> - be event based

Reacting to events?  Generating RDF/SAX events from parsing?

> - must support signatures, digests

Still some consensus need here about how signing RDF is done.  Here
we really have to be precise.

> - possibly have a SOAP interface

SOAP interface - to what level?  Do you mean a general RPC interface,
maybe in SOAP, CORBA, XML-RPC, K-Parts, ...?

> - be easy to use and understand for the programmer

I humbly suggest Redland satisfies many of these [can't comment on
the last one].  I'm still coding and open to suggestions and
especially help!

Dave

Received on Sunday, 12 November 2000 05:57:57 UTC