W3C home > Mailing lists > Public > public-rdf-ruby@w3.org > June 2010

Re: RDFa parser for RDF.rb

From: John Fieber <jrf@ursamaris.org>
Date: Thu, 3 Jun 2010 10:45:25 -0700
CC: "Hellekin O. Wolf" <hellekin@cepheide.org>, Nicholas Humfrey <nicholas.humfrey@bbc.co.uk>, "public-rdf-ruby@w3.org" <public-rdf-ruby@w3.org>
Message-ID: <895F5062-06BC-4C1E-BF2C-5F0335966DB4@ursamaris.org>
To: Gregg Kellogg <gregg@kellogg-assoc.com>
On Jun 3, 2010, at 10:07 AM, Gregg Kellogg wrote:

> RDF::RDFa is a native ruby implementation of RDFa. It's currenly on track with work on RDFa 1.1 based on the latest editors draft. rdf-raptor relies on the Redland/Raptor native implementation, which has more complex installation dependencies. RDF::RDFa only relies on Nokogiri, which is based on LibXML, a standard include in Ruby distributions.

Nokogiri is a C extension wrapper for libxml2 which is not, to my knowledge, a standard Ruby component or dependency.  It is true, though, that libxml2 is much closer to libc end of the you-probably-already-have-it spectrum than libraptor is.  Digging through Nokogiri, I see[1] it has ffi bindings to libxml2, but it uses the C extension by default.  Does anyone know if the ffi route has feature parity with the C extension?  If not, this is a non-starter for jruby folks.

In any case, with a prominent third party C extension dependency, I don't think you can claim a native ruby implementation any more than rdf-raptor can.  Personally, I don't really care whether it is or isn't, but the claim and the reality ought to be in sync.

-john

[1] Or saw, when I switched some time ago from libxml-ruby to nokogiri because the former will happily sefgault if, say, the GC happens to run.  Just set GC.stress = true and watch it explode with even the simplest operations!
Received on Thursday, 3 June 2010 17:46:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 3 June 2010 17:46:08 GMT