Skip to Content.

cat-users - Re: [[cat-users]] Feature request for login name

cat-users AT lists.geant.org

Subject: The mailing list for users of the eduroam Configuration Assistant Tool (CAT)

List archive


Re: [[cat-users]] Feature request for login name


Chronological Thread 
  • From: Zenon Mousmoulas <zmousm AT noc.grnet.gr>
  • To: cat-users AT lists.geant.org
  • Subject: Re: [[cat-users]] Feature request for login name
  • Date: Tue, 05 Apr 2016 14:21:34 +0300
  • Authentication-results: prod-mail.geant.net (amavisd-new); dkim=pass header.i= AT noc.grnet.gr

Hi Tomasz,

On 2016-04-05 13:52, Tomasz Wolniewicz wrote:
W dniu 2016-04-05 o 11:59, Zenon Mousmoulas pisze:
brief answers to your questions:

I see your point about inner identity. Making this configurable as an
on/off option next to anonymity support (general profile properties)
would be OK. I wouldn't even care if it defaults to enabled, as long
as the option is there. It would be fine to just suggest (in a label
next to the actual text input) something like
username@realm-configured-on-profile. No pre-fill or enforcement, I
believe we can revisit this later if necessary.
I wonder if people are aware that in CAT itself you can set additional
hints which are displayed in the GUI before the download is started.
For NCU this hint says "For eduroam use the same username and password
as for USOS" (USOS is our study support system).
Perhaps, if people used this more then the problem would go away?

Yes, but that option is "hidden" under device-specific or eap-type-specific options. You also need to set it per language. And this comes at a different point in time, not when the user is prompted to enter their username. Which prompts for a different question: Do the android app users get to see this text when they pick a profile from within the helper app?

In any case, I am not a UX person, so if this is how/when you think users should get prompted, I wouldn't object to that either. I understand that is a more flexible option, but it still feels different from the sort of hint we are discussing.

As for realms: On the CAT profile an exact (not a suffix nor a
wildcard) realm is configured so far, as that is used for
reach-ability checks (the only visible use AFAIK). If the semantics
were to change, I would prefer itif CAT would support a list of realms
associated with a profile, as I think we are trying to do away with
suffixes and wildcards in general (in an eduroam db context). In such
a case the label/descriptive text should say something like @<realm>
where <realm> can be one of: ...
The realm is used to set the outer identity for methods that use one.
If you do not set the realm then the anonymous identity option is not
available. Thus is the *main* use of the realm, reachability tests are
just an ad-on utility. This is why we cannot have a list of realms or
wildcards there. If I correctly understand your need for lists or
wildcards this would be for sanity testing of user input, I would
consider this to be an overkill. For sanity testing I would for suffix
testing only based on the current realm setting, with an option to turn
this all altogether.

You are right, I totally missed that. The last proposal about sanity checking sounds good as well.



Archive powered by MHonArc 2.6.19.

Top of Page