A lot of people have been doing a lot of talking. There is rough consensus around a short term plan for the client implementation team. I've worked with nick, chris, mfinkle, rnewman and others to figure all this out (they did all the work, I just type). Here's an attempt to distill that, blow me up publicly if I say something you can't live with.
What's the target? Milestone 1 - Firefox Accounts delegating to existing Sync 1.1 account We are going to take the mocks we have and build a fully functional subset of MVP. This is going to have the following properties: 1. implements UX mocks for Firefox Account setup and management on desktop and Android (no FirefoxOS for milestone 1) 2. implements Firefox Account password-based flow against picl-idp.org described at (need a git revision or other stake in the ground here) backend comfort by mark mayo ala rfk are pre-requisites to shipping? 3. invisibly creates a Sync account to do the actual Syncing (so no major desktop code changes needed) (this might require using a stage Sync server to avoid captchas) 4. stores data using the current Sync implementation (does not change the existing sync codebase) (storage architecture 1.1) 5. scope work required to avoid future flag days (but implementation not part of milestone 1) Who's doing the work: picl-idp: dcoates is the lead, with support from rfk, zaach, and dev/ops to be determined Android: Nick Alexander is the lead. Desktop: Lloyd Hilaiel is the lead in constant contact with Gavin, Dolske, and ttalbert Future storage scoping work continues in parallel by: ckarlof working closely with rnewman & warner On android and desktop sides we'll pull in additional help as needed. What's next: First step is to define the steps to get to milestone one. That happens below. Second step is for Nick Alexander to break this down into granular bugzilla bugs for implementation on the android side (happening today). Third step is for lloyd to map these bugs over to desktop. (happening tomorrow) Fourth step is to knock them down one by one, with code landing in ELM. Key Steps: 1a. On Desktop implement HTML based authentication - sign up/sign in 1b. On Android, implement native Account and native UI authentication flow 2. On both platforms, use Sync 1.1 user API (http://docs.services.mozilla.com/reg/apis.html) to create a "derived" Sync user account with username, password, and encryption keys derived from the Firefox Account settings. [1] 3. On both platforms, use derived Sync account to store data into existing Sync 1.1 Thoughts? lloyd [1] On Android, this will look weird -- there will be a Firefox Account and a Sync Account at the same time. This is temporary. If we really care to make this look slick, we can make the delegated Sync account go away. On Desktop, there will be the usual evidence of Sync. _______________________________________________ Sync-dev mailing list [email protected] https://mail.mozilla.org/listinfo/sync-dev
