- From: Dan Brickley <danbri@w3.org>
- Date: Wed, 29 Sep 2004 10:30:14 -0400
- To: Christopher Schmidt <crschmidt@crschmidt.net>
- Cc: john.nj.davies@bt.com, Libby.Miller@bristol.ac.uk, Kianoush.Eshaghi@metadat.at, charles@w3.org, www-rdf-interest@w3.org
* Christopher Schmidt <crschmidt@crschmidt.net> [2004-09-29 10:22-0400] > On Wed, Sep 29, 2004 at 01:49:51PM +0100, john.nj.davies@bt.com wrote: > > > > There is also the mbox_sha1sum property which uses a hash of your email > > address to uniquely identify you if you do not wish to disclose your > > email address. > > > > <foaf:mbox_sha1sum> > > 3edg453144de22607a169e585c487967d035cd27 > > </foaf:mbox_sha1sum> > > As with all literals, you should keep in mind that adding newlines like > this significantly changes the value of the literal, which is a bad thing > with terms like this where the sha1sum is used as a unique identifier: > > <mbox_sha1sum>2388d8d8d</mbox_sha1sum> > > is different, and will not smush with, > > <mbox_sha1sum> > 2388d8d8d > </mbox_sha1sum> > > Common practice is to include only the 40 character hash, and not any > whitespace or newlines. It was recently (foafcamp and foaf galway meetings) re-drawn to my attention that language tagging also affects plain literal comparisons. The FOAF spec should probably be updated to encourage people not to language tag mbox_sha1sum, and to note some extra conventions for identity reasoning here (ie. ignoring lang tagging) which go beyond what generic tools will be able to accomplish. Dan
Received on Wednesday, 29 September 2004 14:30:15 UTC