shibboleth-dev - Re: Java SP Entrance Point
Subject: Shibboleth Developers
List archive
- From: "Spencer W. Thomas" <>
- To:
- Subject: Re: Java SP Entrance Point
- Date: Mon, 02 Oct 2006 16:47:11 -0400
- Organization: JSTOR
Chiming in here, a little late, as a SP developer.
Our authn/z uses the filter approach (in Java -- our perl definitely
follows the programmatic approach for authz, with some authn mixed in
for the IP case.) The filter approach provides the cleanest separation
, and allows development of simple apps (for which authz is either "yes"
or "no") to be totally innocent of any authn/z knowledge. For these
apps, the filter denies access if the user is not authz. Even
authz-aware apps don't have to deal with the unauthz case, and can focus
on application-specific authz processing.
Where an API is good, is in providing a way of getting at the
information that the filter has deposited in the request/session. It
can be pretty thin, just hiding the attribute names, or it can be
thicker where there is a need for commonality of function in the client.
=Spencer
- Re: Java SP Entrance Point, Spencer W. Thomas, 10/02/2006
Archive powered by MHonArc 2.6.16.