[foaf-dev] @deprecated mbox? (and phone? etc)
danbrickley at gmail.com
Fri Feb 1 16:33:34 GMT 2008
On 01/02/2008, KANZAKI Masahide <mkanzaki at gmail.com> wrote:
> Hi, tracking the active mail box is nice and important idea.
> I wonder using foaf:pref_mbox (or whatever, like skos:prefLabel) to
> specify the current active (or preferred) mail box. Wouldn't it be
> simpler if you change active mbox several times ?
I wondered about this (and about SKOS too :) ...
The fact that I went from daniel.brickley at bristol.ac.uk to
danbri at w3.org to danbri at danbri.org is probably not very interesting to
the world. For now, we can already query (like in my example) for
mailboxes that don't have an outgoing "isReplacedBy" property. But
yes, I see there is some risk that FOAF docs will fill up with old
mailboxes, and each will be flagged foaf:mbox, so apps looking for
current address may have more work to do (ie. more complex queries).
We can't change foaf:mbox much, having marked it as stable, but I am
interested in having a design for current/active and work/non-work
flags that works across IM, email, phone and Web. The dct:isReplacedBy
design I think is OK for now, and gives us time to have that
discussion while deploying something simple today that doesn't need
new terms. Does that make any sense?
We could perhaps also have a deployment convention that said, "there's
no need to list old mailboxes in your main foaf description as
foaf:mbox, although you can list the fact that they are replaced by
newer ones. That would mean that any identity reasoning based on those
mailboxes wouldn't work any more, ... unless they were using the
corresponding foaf:mbox_sha1sum instead. This is worth encouraging
anyway. I've been talking with Libby Miller and Dirk-Willem van Gulik
btw about extending the definition of foaf:mbox_sha1sum to include
(within the value string) an optional nonce
(http://en.wikipedia.org/wiki/Cryptographic_nonce), to make them more
private. But more on that separately...
More information about the foaf-dev