COmanage Developers List

[Benn Oshrin <>]

On 10/27/11 9:13 PM, Scott Koranda wrote:

> > > Need to consider what happens if the sponsor becomes invalid.
>
> Can or should we limit the notion of 'sponsor' simply to
> enrollment?

There may be two use cases here: sponsor to get enrolled and sponsor to 
stay enrolled. Regardless of LIGO requirements, COmanage should probably 
support both.

Marie:
>> Sponsors probably don't move around/leave as often as the people they
>> sponsor, but it seems like a good idea to require manual removal of
>> people whose sponsors leave.  If A sponsors B and years later, B
>> sponsors C, if A leaves, I don't think old data should affect B nor
>> cascade to C unless we build in a way for that data to expire.

The difference might be as simple as making expiration optionally 
dependent on having a valid sponsor. If your sponsor goes away, then the 
expiration/notification process gets triggered (until, presumably, 
somebody finds a new sponsor for you). (We should obviously document 
this eventually.)

> > Flagging invalid sponsors in someone's dashboard might be useful but
> > otherwise shouldn't a difference so long as the person being sponsored
> > isn't still pending.

Agreed.

> > How many levels of sponsorship do we need?  I added a sponsor for
> > membership into a group in cm_co_group_members here:
> > https://spaces.internet2.edu/display/COmanage/cm_co_group_members
> >
> > Is that level sufficient or will there ever be a need for sponsors
> > into the CO or CMP level?

I think sponsor gets attached to co_person_role. Alternately, if there 
are multiple sponsors per person (which I hope there aren't), we need a 
co_sponsors table (but let's try not to go this route).

I hadn't thought about sponsors as part of group memberships... I wonder 
if CORE (COmanage REgistry) should do this or if we should rely on 
Grouper (or the GMS) to worry about it. The CORE design for groups was 
always intended to be pretty minimal. (Can Grouper track sponsors for 
group memberships? Is there even a LIGO requirement around this?)

-Benn-