InCommon Technical Discussions

["Cantor, Scott" <>]

> I do like the idea of some commonly-available "attribute diff" tool that 
> could
> make discovering this change easier. Obviously, it is something that you can
> figure out through careful testing during the IdP upgrade, but it would be
> fairly easy to overlook if you didn't specifically search for it.

I think we added something to our upgrade instructions about this, but since 
it requires getting a sample of an ID from before the upgrade to compare to, 
it makes it difficult to do much beyond documenting it. Even a test service 
is 
only helpful insofar as you can get somebody to poke it pre-update.

All in all, the real fix is "end use of pairwise ID for research sites" to 
the 
extent possible.

As long as people insist on upgrading by doing new installs, against our 
advice, this is just going to keep happening for anybody that uses generated 
IDs. 4.0, etc. No way to stop it. Not caring is the fundamental issue, and I 
can't fault any SP for basing their decisions on what people care about.

-- Scott

Attachment: smime.p7s
Description: S/MIME cryptographic signature