CloudServices/Meetings/2009-07-13
From MozillaWiki
< CloudServices | Meetings
Agenda
- Migration plan for usernames (anant)
- Snapshot release plan (mconnor)
- Server scalability update (Toby)
- 1.0 PRD next steps (Ragavan)
- Blog post ideas:
- Changes to server API (Toby)
- Username migration - why, how details (Anant)
- Recent code improvements (Mardak)
- 0.5 release (Connor)
- Project update (Ragavan) Done
- Crypto details, password/passphrase idea (Dan)
- Ideas for onsite, sprints? (All)
- User API (Toby)
Carry-over:
- AI: Dan to put together a wiki page explaining how users can turn on verbose debugging.
- Dan has a page that's almost done. Yay!
- AI: Dan is going to write a first draft of a WEP for Add-Ons sync/storage API (Ragavan: medium priority).
- MConnor is going to work on the benchmarking after the unit tests going. He is going to open a bug to do this. We need to measure sync time and memory usage. Dan and MConnor will basically figure out what needs to be measured to help answer whether or not we are going to need a crypto proxy for fennec.
- Dan is going to ping mconnor as this is very high priority.
Notes:
Migration plan for usernames
- Usernames are case sensitive though it wasn't our intention.
- New rego script so users' DN will be stored in lowercase in LDAP.
- Update mysql tables to lowercase all usernames.
- LDAP update.
- Server code update.
- This will need downtime, but should be user transparent.
- When: Saturday (July 18) Communication: Today (anant) Who: Toby/anant/zandr
Snapshot release plan
- Proposal to ship weekly if we are on a 4-week cycle.
- If we can handle wipe-out of user data on upgrade (which sounds like we can), we should do this.
- AI: Mardak is going to file a bug on potential improvements.
- We'll plan on doing weekly releases on Wednesdays.
- Should we have a global cap on # of items, per data type? May not make sense for all engines (like bookmarks).
Server scalability update
- Hardware is available (blades), zandr is going to set it up. We need 3-4 machines just to replicate current load.
1.0 PRD next steps
- Identity will be experimentally supported. Remove content munging until we can get the UX figured out.
Ideas for onsite, sprints?
- Bloggable is a good metric.
- Don't have to be just experimental features, can also be stuff for 1.0.
User API
- Just about ready. Implements the User API spec.
- We should use the JEP style format for future.
Other
- How to address captcha issues during user registration?
- Implement recaptcha proxy for mozilla? Not really needed, 'cos the most troublesome one is NoScript.
- Options: Detect NoScript and behave accordingly?