Re: [apps-discuss] Webfinger / SWD Issue #3: direct versus indirect discovery

[Blaine Cook <[e-mail]>]

*sigh*

This is crazy. There are a half-dozen of us debating theoretical ways to
structure a thing that may or may not ever get used, so that some specs can
get finalised so that *maybe* Microsoft and Google can ship a new version
of OpenID (that doesn't adequately address one of the core usability
shortcomings of the first two OpenIDs).

I don't know if Evan Prodromou or anyone from Diaspora or anyone else who
has *actually tried to deploy* similar things is watching this thread, but
if they are, I hope they pipe up, because there is absolutely zero
consideration for actual implementations happening in this discussion.

Also, keep in mind that if *we're* confused about the host-meta processing
rules, or if my "single-code-path" arguments are being dismissed because
"it's easy either way", then we're in a lot of trouble. Refer to the fact
that we've switched from XML to JSON because XML is too hard for most
developers.

The point of all this is to make it accessible and easy, since we have a
chicken-and-egg problem. If this is going to work, it's going to take the
whole long tail, not just Microsoft and Google. The specs we have right now
clearly don't work, but idle speculation on our part isn't going to make
any of this work.

b.