OIDC Deployment Working Group

[Nathan Dors <>]

Thanks for the info.

Sounds like pyoidc and JWTConnect-Python are both reasonable choices at this point.

(Apparently, cc'ing the list doesn't deliver to the list, so I'm shifting this thread to the list.)

On Wed, Mar 13, 2019 at 4:22 AM Roland Hedberg <> wrote:

> 12 mars 2019 kl. 19:14 skrev Nathan Dors <>:
>
> Roland, got your ears on?

:-)

> On today's call we discussed the deployer's need to identify popular and conformant OIDC relying party libraries for internal testing and also to nudge customers toward good libraries.
>
> What's the current state of the jwtconnect.io libraries, pyoidc vs oidcrp projects, and other relevant news related to relying party libraries?

The Python and the Java implementations are complete. Needs testing by real users and better documentation.
The _javascript_ implementation needs some work.

The finance side of maintaining the JWTConnect libraries are not finalized.
Will take another 2-3 month at least. If everything comes through we’ll have a nice bit of funding for the next 5 years or so.

The JWTConnect libraries are the new kids on the block and therefor hasn’t been tested the way that for instance pyoidc has.
Also pyoidc has a user community. There are users of the JWTConnect-Python libraries but I wouldn’t call it a community yet.

What speaks for the JWTConnect libraries (in my mind at least) is:
- that they are designed based on the experience gained from writing and maintaining for instance pyoidc.
- that IdentityPython (https://idpy.org) is going the JWTConnect-Python way.
   There will for instance be new front- and backends to SATOSA built on JWTConnect-Python.
- that there is a pretty high probability there will actually be money to pay people to maintaining the libraries

—Roland
The reward for conformity was that everyone liked you except yourself. -Rita Mae Brown, writer (b. 28 Nov 1944)