Re: [DotGNU]Using FOAF for Virtual Identities

[Peter Minten]

address@hidden wrote:
> 
> Peter,
> 
> As I'm fond of pointing out, MACS can be made compatible with any
> identity standard. Internally, it implements a decent standard that
> stands on its own, but whose great usefulness lies in being easily
> mappable. That is, easy to make compatible with other standards.

OK, that solves the problem of can x (where x is a VI standard) be used with
MACS.

> As such, it's neither FOAF, nor Liberty Alliance, nor Passport, nor
> Genio; yet, it can become all of them. MACS calls these other
> standards "Methods" and it's been pretty easy to map strange and
> wonderful Methods for use by MACS. For example, the Yahoo! method
> can log you in (using your Yahoo! username and password, of course.)
> With a little more HTML screen-scraping, a Profile Method Client can
> be written to let MACS know, e.g., how many unread emails you have
> at your Yahoo! account.
>
> Being unfamiliar with FOAF, I'll introduce the concept of Frontends.
> Where Methods let MACS use random sources of user/account/identity
> data, Frontends let MACS deliver decisions and answer queries in
> random formats. There are Frontends for SOAP and XML-RPC at the
> moment. A Jabber Frontend is currently on the back burner.

The mysterious MACS explained :-). So methods are other auth protocols, PMC's
work on methods and front-ends are used to communicate with other
programs/persons.

> So, depending on what exactly FOAF is, and how you want to use it,
> a set of Methods (one each for authentication, authorization, and
> profiles,) and/or a Frontend would be written. This code would not
> need to know too much about how MACS works, but it would have to
> know a lot about FOAF. (That's the point -- this code would embody
> MACS' knowledge about FOAF.) Then snap your fingers.
> 
> Voila! MACS is now a FOAF service. (Or a reasonable facsimile. ;-)
> 
> The interesting thing that needs to be added to MACS to make it
> useful for DotGNU is what FrePort wanted to do. That is, keep
> profile data secret unless the user explicitly said otherwise. At
> the moment, MACS has no such facilities or guarantees. The format,
> protocol, &etc. are not irrelevent questions, but they are moot.

The FrePort stuff is important, do you know how John wanted to achieve that?