+cc foaf-dev On Thu, Dec 17, 2009 at 3:19 PM, Brion Vibber <[email protected]> wrote: > On 12/16/09 10:20 AM, Toby Inkster wrote: >> >> The FOAF spec was updated yesterday. The only change that affect's >> StatusNet is that foaf:holdsAccount was deprecated in favour of >> foaf:account. >> >> The news: >> http://lists.foaf-project.org/pipermail/foaf-dev/2009-December/009903.html >> >> Quick patch to bring StatusNet up to date: >> http://buzzword.org.uk/2009/statusnet-0.9.x-foafupdate.diff > > Is it best to drop the holdsAccount entirely, or provide both entries for > compatibility with any clients that might already be using this information?
The 'best practice' for these kinds of transition is still evolving. I should also emphasise (and Libby also mentioned) that we marked 'holdsAccount' as being 'archaic' rather than 'deprecated', in acknowledgement that existing usage remains out there and that we can't force people to migrate; sometimes it is impossible or inappropriate to update certain code or data. There is still some openness around how exactly to deploy the property; in particular whether the URIs for each OnlineAccount are simply the URIs like http://identi.ca/danbri/ ... this is my favoured design, but this point has not been pushed in the latest revision. It would be reasonable to hold off code / data changes until this point is confirmed, since StatusNet uses a slightly different idiom currently. For StatusNet, I would suggest ideally exposing both properties for a limited time, if that is acceptable to you in terms of code complexity and bandwidth use. Then we'll work to get client / consumer code updated, and switch over completely in a few months. cheers, Dan _______________________________________________ StatusNet-dev mailing list [email protected] http://lists.status.net/mailman/listinfo/statusnet-dev
