Benjamin Trott ben at
Wed May 25 08:39:26 PDT 2005


Yes, I'm absolutely happy to kill of the Crypt::OpenPGP requirement--I
didn't even think it was required, I thought it was dynamically loaded. I
may be wrong.

As for RDF::Core, it's pure Perl, and depends only on XML::Parser. It does
have a rather annoying warning about a variable already having been declared
via my (twice in the same scope), but other than that, it's pretty


On 5/25/05 5:17 AM, "Christopher Schmidt" <crschmidt at> wrote:

> On Wed, May 25, 2005 at 02:39:12AM -0700, Brent 'Dax' Royal-Gordon wrote:
>> Brad Fitzpatrick <brad at> wrote:
>>> Mark Smith and I were discussing today a new module today, tentatively
>>> called Net::OpenID::UserProfile, which would take a VerifiedIdentity
>>> object from Net::OpenID::Consumer and use the foaf/maker/rss/atom data
>>> that it discovers and makes available, and then actually goes and crawls
>>> those resources (perhaps in some defined order of preference) to build a
>>> UserProfile object with things like email (or hash), name, nick, blog URL,
>>> etc.
>> Unfortunately, XML::FOAF is a really heavy module; it depends on
>> Crypt::OpenPGP, which pulls in about twenty crypto modules, many of
>> which don't seem to install cleanly via the CPANPLUS shell.  (They
>> work fine with standard for some reason.)
> Given the current state of FOAF signing, it's probably best to just kill
> the Crypto stuff out of it: I have a collection of ~70,000 RDF documents,
> including multiple thousands of which are FOAF, and 7 (yes, 7) are
> signed. However, that still leaves the module with the very heavy
> RDF::Core dependancy. Unless someone can tell me that my perception that
> this is a blocker is just myth, which is totally possible, in which case
> I can get into tearing apart XML::FOAF and making into what it needs to
> be.

More information about the yadis mailing list