Thanks for the feedback Ken, very useful. Would be good to catch up over a 
coffee or something before you leave.

A few quick comments below, but will answer more substantially when I start to 
cut the next draft and look at your points in more detail.

> Rhys,
>   I've done my homework (fearing Leif's steely eye) and reviewed the draft on 
> the UI considerations (Sept 25 version). I'll be at IETF Sunday night through 
> Wednesday, in case you'd like to chat further about this, but have to leave 
> before the ABFAB session on Thursday. 
>   There's a lot of "still to do" sections in the draft, as you know – I'd be 
> glad to look at them.
>   A few overarching comments, and then some specifics on the draft.
>       1. I'm glad that you're looking at the larger discovery problem issues 
> rather than just the ABFAB use case. That said, it does bring the discussion 
> close to other efforts such as AccountChooser. In speaking to John Bradley 
> recently, they have many of the concerns that you mention (most notably they 
> are now considering how to control what IdP's can register with an SP or an 
> accountchooser location – trying to get into a dynamic process such as the 
> one you mention, etc. Some reference to those discussions might be useful.
>       2. There is almost no mention of privacy related issues. As a matter of 
> form, there should be such as a section. Maybe mention the issues around 
> hiding IdP's (we've had instances where people wanted to conceal their IdP 
> choices from interested governments –certainly not the US :)) - to whether an 
> IdP follows privacy principles (via an end-entity tag or some other 
> information that might want to be shown to the user.) One of the other ways 
> this could be brought out is via some mention of identifiers.

Absolutely, given I’m one of the authors of 6973, not having a privacy section 
just wouldn’t be right :-).


>   In your terminology section, I would introduce the identifier vs identity 
> distinction, one I'm trying to hammer on in NSTIC. You mention that a user 
> MAY have multiple identities. It would be good to say that an identity MAY 
> have multiple identifiers associated with it, in order to preserve privacy, 
> etc. 

Good point, agree totally.


>   The first sentence in section 4 could use some smoothing – seems like an 
> awkward sentence. In that same section, you talk about a headless mode 
> operation, but don't define it (and I, as a semi-knowledgeable reader don't 
> know what it means.)

OK I’ll try to sort that out. FYI, headless mode operation would be - how do 
you do authentication in a world without a GUI? (e.g. machine to machine 
authentication via ABFAB). 



>   The first sentence in section 6 uses "firstly" - not English on this side 
> of the pond. And at the end of section 6.1, a typo (exmaple instead of 
> example).
>    In section 7, first sentence needs the word "tell" to be singular (tells). 
> Also, the many-many scenario you talk about, in terms of complexities, might 
> be best addressed with a single identity and a mechanisms for that identity 
> to convey the role its in. Maybe hard in an ABFAB case, but maybe not and 
> worth a mention as an alternative.
>   Again, willing to review new sections as they are added. A topic close to 
> home for me.

Super.

Rhys.
--
Dr Rhys Smith
Identity, Access, and Middleware Specialist
Cardiff University & Janet - the UK's research and education network

email: [email protected] / [email protected]
GPG: 0xDE2F024C

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab

Reply via email to