An open discussion list for topics related to the eduGAIN interfederation service.

[Martin Matthiesen <martin.matthiesen AT csc.fi>]

Hello,

I think we are again reaching the limits of email. I feel I have made my 
point of view known and I can only hope that the concrete use case and its 
problems spark off some changes in the way attribute semantics are handled 
within eduGAIN and within the national federations. Please find my answers to 
Leif's questions below.

----- Original Message -----
> From: "Leif Johansson" <leifj AT sunet.se>
> To: edugain-discuss AT geant.net
> Sent: Monday, 8 December, 2014 20:45:06
> Subject: Re: [eduGAIN-discuss] eduGAIN and non "academic" IdPs

[...]
> 
> I believe this is the correct reference for epa:
> https://spaces.internet2.edu/display/macedir/eduPersonAffiliation-2011-draft-02

I knew this reference, but found it too vague to refer to it. It does not 
even define "faculty", only mentions it. That's why I referred to Feide. I 
find the top of Swamid's writeup (linked below) even clearer, but it is in 
Swedish.

[...] 
> Just out of curiosity: did that professor contact his IT department and
> if so what was the outcome of that discussion?

I don't know. 
 
[...]

> 1. Why do you think 'employee', 'member' and 'student' is almost
> universally available while 'faculty' and 'staff' is not?

I guess one reason is this:
https://portal.nordu.net/pages/viewpage.action?pageId=31197805

And with that, Swamid doesn't even completely break the standard. It just 
makes it impossible to just use faculty for countries where Academic would be 
just "faculty".

> 2. What do you think SWAMID (in particular) did in order to get those 3
> almost universally supported by IdPs in SWAMID?

Likely a lot. And part of my motivation to argue here is that that work does 
not go to waste.
 
> 3. What power does SWAMID have over its customers (its a trick question)?

Apparently not a lot. I just don't believe that model works in this context. 
And I am fully aware that changing it is hard. I just don't see an 
alternative. As I wrote earlier: Defining standards without the possibility 
to get them implemented, seems fruitless. And I don't want to single out 
Swamid, it just so happens that my service has Swedish data and it would 
therefore be nice if it could be easily accessed in Sweden.

Martin