[sugar] Obtaining Buddy objects as contacts are encountered
Mon May 14 07:55:26 EDT 2007
On Mon, 2007-05-14 at 11:24 +0100, Simon McVittie wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> On Fri, 11 May 2007 at 17:24:49 -0400, Dan Williams wrote:
> > On Fri, 2007-05-11 at 19:25 +0100, Simon McVittie wrote:
> > > The code only seems to have 'nick', which is set to the name you enter
> > > when you first switch on a new OLPC. Is your position that there should
> > > be separate attributes, 'nick' which you can change, and 'name' which
> > > you can't? Is there any design in which this is documented? In the
> > > absence of any particular reference, I'd been assuming the code
> > > implements the design.
> > Yeah, we should do this. I assume that Vcard supports First/Last name?
> > Note that we also get into issues then with ordering/localization,
> > because some locales (Hungary, for example) use family name _first_.
> > But that's a problem for sugar, really.
> vCard supports both FN (Formatted Name = display name, e.g. "Dan
> Williams" and N (structured Name fields, e.g. "Williams;Dan;;Mr.;"). The
> fields of N include family and given name, rather than first and last
> name; there appears to be no way to indicate which comes first in the
> locale of the subject of the vCard.
> Telepathy's full Contact Info interface is under revision - the one
> currently in the spec is far from ideal, but the replacement needs
> further discussion - so I'd suggest we add "full-name" to the OLPC buddy
> properties interface. We could always replicate one to the other
> automatically, at some point.
> For Salut, I think we should also add "jid" to the OLPC buddy properties
> (it ought to be in contact info, but again, we don't have a good
> interface for that), to make it easier to link mesh and server
> identities. In my proposed implementation (which I'm still writing!) we
> need to use the JID to link identities, because retrieving the public
> key takes network round-trips, and it's problematic to have people turn
> up and start interacting in a tube before we actually know who they are.
> I'll add full-name support as part of my current round of API changes,
> if there are no objections.
This all sounds good.
> Are we happy for the child's full name to be public (to anyone who knows
> their JID), from a privacy point of view? I realise the answer is
> probably "yes", but I feel I should ask...
> What's the intended UI for this? I assume we should use the name entered on
> first boot to populate both the full name and the nickname, then let the child
> change their nickname to something else later?
> Once again, if there's any design documentation I should be consulting
> on this, please let me know. Otherwise I'll just carry on trying to make
> reasonable decisions so we can get something implemented.
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
> Comment: OpenPGP key: http://www.pseudorandom.co.uk/2003/contact/ or pgp.net
> -----END PGP SIGNATURE-----
> Sugar mailing list
> Sugar at laptop.org
More information about the Sugar-devel