Just to clarify, xauth will be available to mobile applications (who apply) going forward to authenticate users, not just a one time way to exchange stored usernames and passwords?
On Feb 11, 10:18 pm, Raffi Krikorian <ra...@twitter.com> wrote: > hi all. > > this is a long overdue e-mail, but i wanted to tease out some of the > directions that Twitter is going with OAuth. i want to touch upon four > topics: delegation, OAuth WRAP/2.0, username/password OAuth token exchange, > and basic authentication deprecation. > > *DELEGATION - OAuth Echo* > > twitter users love posting media on third-party sites, and delegation in > identity verification is one of the major hurdles for an OAuth-enabled > twitter client to succeed. i started a series of blog posts around the > following problem: > > You're an OAuth enabled Twitter client, and you've already authorized your > > > user. Your user wants to use a media providing service like TwitPic. > > TwitPic, currently, asks for the username and password of your user so it > > can store the photo on behalf of the Twitter user. You don't have that > > username and password, so how do you give the ability to TwitPic to verify > > the identity of your user? > > check out the proposal for what we're calling "OAuth Echo" > athttp://mehack.com/OAuth-echo-delegation-in-identity-verificatio. please > feel free to comment there, or on the twitter development talk mailing > list<http://groups.google.com/group/twitter-development-talk>(or, even > just reach out to me directly). i think this experiment in > engaging the community around designing this security/identity workflow has > been definitely a success, and i feel we're rapidly converging on a solution > for identity verification delegation. in parallel, we're going to start the > process to engage our media providers in the conversation as well, and we're > hopeful we can move this forward quickly. > > *OAUTH WRAP/2.0* > > OAuth is evolving, and we at Twitter are keeping up with it. that being > said, we're keeping our eyes on OAuth WRAP and OAuth > 2.0<http://wiki.oauth.net/OAuth-WRAP>. > we like a lot about it: > > - it requires the use of SSL; > - there is no custom signing mechanism -- you simply pass us a token, and > that token is secured by SSL; and > - it formalizes a bunch of "profiles" that we've been actively thinking > about (e.g. a username/password exchange) > > in general, we really like WRAP/2.0 because it's just *so* easy to implement > from the client side. there are no longer questions around creating the > proper signature base string, etc. we're sure that developers will like it > as well. we've started work on an internal implementation of OAuth WRAP and > we envision that we'll simultaneously support both OAuth 1.0a and WRAP/2.0 > for a while. our hope is to get WRAP out the door soon (and before we > finally deprecate basic authentication). > > *USERNAME/PASSWORD TO OAUTH TOKEN EXCHANGE - xAuth* > > @rsarver and @noradio announced that we are going to support a mechanism > where a username and a password can be directly exchanged for an OAuth token > and secret -- we're calling this xAuth. if you've been watching the mailing > list, Seesmic Look <http://seesmic.com/look> has been a beta partner in > testing xAuth exchange (and @abraham has already detailed how it > works<http://the.hackerconundrum.com/2010/02/sneak-peek-at-twitters-browser...>). > > because we're moving everybody off basic authentication, we originally > envisioned this as a mechanism for developers to exchange all the username > and passwords they have in their databases for OAuth tokens en masse. > that's still one of our use cases. another use case is around environments > where software can't bring up a web browser (e.g. set top boxes, game > consoles, embedded devices). we want to support those as well. > > you're going to have to apply to get access to this exchange mechanism (by > sending e-mail to a...@twitter.com), but, in general, all applications except > web applications will get access. we feel that the xAuth exchange allows > for the best mix of security and user experience for desktop and possibly > mobile applications. web applications will simply have to use OAuth as it > was designed, and send their users through the web flow. > > *BASIC AUTHENTICATION DEPRECATION* > > yup - it's still happening. we're targeting June 2010. everybody, > including legacy applications, will have to move over. > > for those who are building new applications, use OAuth. save yourself the > transition time later, and start thinking about it now. for those who have > applications already out there, it would be really beneficial to start > thinking about a migration path right now and we're here to help. if you > need it, please feel free to reach out to us and we'll help you figure out > what you need to do. > > to help entice you over, as you know: > > - we have increased rate limits on api.twitter.com to those who are using > OAuth (350 calls to the REST API per hour -- and increasing towards > 1500/hour); and > - (as some of you are painfully aware) you can only set a source > parameter with OAuth calls to status/update. > > we know some of you think there are hurdles in places to converting over to > OAuth -- suffice it to say, we're actively trying to address them. some > potential hurdles are mentioned in this e-mail, and if you think there are > more then please feel free to reach out to us. again, we really want to > make this switch-over easy for everybody. > > thanks! and, as always - feel free to reach out to us anytime here, or to > @twitterapi. > > -- > Raffi Krikorian > Twitter Platform Teamhttp://twitter.com/raffi