[uf-discuss] XFN Relationship types

[Dan Brickley]

Julian Bond wrote:
> I'm seeing a lot of XFN that uses rel="contact". But there's also 
> quite a few cases where rel="some other relationship" is being used 
> along with other qualifiers.
>
> Is there some standard here we should be encouraging like rel="Contact 
> Acquaintance" or do we expect developers to know that an Acquaintance 
> is also a Contact? And if it's the second is there some mapping table 
> on a wiki somewhere?
A related thing confused me:

Is "X xfn:acquaintance Y" a claim that is considered consistent with the 
weaker claim "X contact Y" ?

Some sites I have buddies listed and since that site isn't so well 
informed about how well I know those people, they'll emit only 'contact' 
links; others have more detail and understand that we have met (and are 
acquaintances) or even that we're friends.

The 'friendship (pick at most one)' rule in http://gmpg.org/xfn/11 can 
be read in two fundamentally different ways. One is that these are 
mutually exclusive states of affairs: call it Reading (A), "If I'm your 
friend, then I'm not a contact or acquaintance". This reading takes the 
constraint to be a rule about the world. Another, reading (B) is that 
"If I claim I'm your friend in some document, I shouldn't also state 
that I'm a contact or acquaintance". This reading takes the constraint 
to be a rule about document syntax / notation. These two styles roughly 
correspond to the comfort zones of RDF vs XML schema languages, btw. RDF 
schemas express generalisations about the world; XML schemas

I've tended towards the latter reading; since it fits with the webby 
model of scattered, partial information. Last.fm might know that we're 
contacts; Flickr might know that we're friends. And missing information 
isn't necessarily the same as being broken.

How often is something like 'rel="Contact Acquaintance"' (ie. multiple 
from the friendship options together) seen in practice? Does anyone have 
a good list of other common values seem alongside the well-known XFN set?

cheers,

Dan

--
http://danbri.org/