Shibboleth Developers

Phone #:  (800) 541-1710
Pin #:  0142203

Agenda:

1) Improving WAYF usability. The PopUp listis getting a little 
long..... are there changes we'd like to make, to improve usability...

2) Shib and Middleware Diagnostics -- From the beginning, we've seen 
many demonstrations of how confusing it is to diagnose problems in a 
distributed system like Shibboleth. When sites are first starting to 
work with Shibboleth, even simple config errors can result in 
confusing messages; worse, different errors can sometimes result in 
the same message. And we're not even sure yet of what kinds of 
messages and problems  may surface when a site institutes large scale 
use of Shib.

A general  approach has been suggested by the E2E MD group. Here's 
the 100K foot view --  have both the origin and target generate log 
files with log entries threaded per transaction. Some sort of process 
would collect origin and target log files and process them, 
discovering the threads. And Help Desks and Sysadmins could look at 
the threads. And yes, there's literally a bazillion really hard 
questions that have to addressed to make  this idea operational.

I'd like to suggest, tho that we begin by discussing how portions of 
the architecture of the Shib implementation might have to be changed 
to work well in  this model. After we get a better understanding of 
the Shb related issues, we'll probably have a joint discussion with 
the E2E MD group.

Some questions that we could explore include:

1) how to implement some sort  of thread thru a set of messages

2) an "architecture for message handling". Do we just generate error 
text like we do now, or should we implement a slightly more 
centralized mechanism for interpreting the error  codes?

... and I'm sure there are lots  more......