OpenID status update

[Martin Atkins]

Brad Fitzpatrick wrote:

> 
> I'm fine with adding an integer.  (effectively the major)  Minor protocol
> revs will be implied by the precense of new parameters.
> 

For what it's worth, I think it's pointless. Everyone will ignore it
until there are multiple versions anyway, so all you're doing is making
the URL a few bytes longer for no real benefit.

As soon as the protocol changes in a backward-incompatible way, we've
lost anyway. There'll still be loads of existing software out there
which won't know how to negotiate versions and won't be able to do the
new version. It's not like TypeKey where there is one server which
supports the old version for the sake of the older consumers; here we
have many ID servers as well, so the problem exists in both directions.

I guess what's being proposed is that the consumer makes a request with
version=2, but then the ID server responds quoting version=1 and the
consumer downgrades. However, that assumes that all consumers will be
able to support all older versions, which I think is a bad assumption
since once all of the "major" identity servers move to the hypothetical
version 2 home-rolled consumers (which don't use an "official" library)
will stop supporting older servers because they won't encounter them
often enough to realise that they must.