grouper-dev - Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes
Subject: Grouper Developers Forum
List archive
Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes
Chronological Thread
- From: "GW Brown, Information Systems and Computing" <>
- To: Tom Barton <>
- Cc:
- Subject: Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes
- Date: Thu, 26 May 2005 15:48:30 +0100
--On 26 May 2005 08:58 -0500 Tom Barton
<>
wrote:
I see what you are getting at. Currently I think that the listVals method returns a Subject multiple times - one for each effective membership and once for an immediate membership. As the membership list returned is not sorted the same subject will generally not be grouped
GW Brown, Information Systems and Computing wrote:
2)Similarly, a GrouperList.firstInChain method would allow me to
indicate in the UI which immediate membership of the current group
caused an effective membership.
_GrouperList.via()_ provides the first group in the via chain.
Confusion over direction! I was working from the 'working' group to the
immediate membership ultimately responsible for the effective membership
i.e. if groupA has groupB as a member has groupC as a member has subject
Z as a member I was interested in groupB->groupC
Algorithmically, it is more efficient to start with a subject and
determine all of its effective memberships than to start with a group and
recursively determine its effective members. The API manages effective
membership using the former approach. To best leverage what's already
there, perhaps what is needed is a method that takes the set of all
"upward" membership chains starting from a subject and prunes that set to
retain only those chains that terminate on a specified group. That
depicts the set of "reasons" why the subject has the effective membership
(or priv).
I think that the listEffVals and listVals methods ought to return a Set - each subject only once. Rather than trying to provide links here to the effective group memberships separately, the subject would be a link which, when clicked, would resolve to either:
Should such a method be in the UI or API?
1) a summary page listing all ways in which the subject is a member of the group and probably some links to effect changes - but this could be quite difficult to explain to end users
2) the priv/membership screen as we do now - if a subject is a member once via an immediate membership
I think it ought to be straightforward to write a query against Grouper_List to return GrouperList objects for each occurrence of a subject as a member of a group - so this should be in the API. For each effective membership I would show the information on chain.png
e.g. can you revoke a privilege for an individual when it was granted
to a group?
You should not be able to. If you can it is a bug. You need to remove
the individual from the group that they are an immediate member of to
revoke the privilege.
I wasn't suggesting that this happened - just that anyone using the UI
would in effect try to do this. I think we need to make it clear, in the
UI, how a subject came to have a privilege/membership - atleast as a
prerequisite to modifying privileges/memberships. It is impossible to
know what to expect otherwise.
See above.
Tom
----------------------
GW Brown, Information Systems and Computing
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, (continued)
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/18/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, Tom Barton, 05/23/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/24/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, blair christensen, 05/24/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, blair christensen, 05/24/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, Tom Barton, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/24/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, blair christensen, 05/24/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, Tom Barton, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, Tom Barton, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/26/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, Tom Barton, 05/23/2005
- Re: [grouper-dev] Grouper design call, Wednesday, 18 May 2005, 1200EDT (1600Z), 60 minutes, GW Brown, Information Systems and Computing, 05/18/2005
Archive powered by MHonArc 2.6.16.