Websites/WebifyMe/Meeting 2011-08-25
From MozillaWiki
Contents
Attendees
- williamR
- grace
- morgamic
- malexis
- stas
- krupa
- stephen
- chrissie
Agenda
Intro:
- Provide general overview and recap of the project.
Planning:
- Were there clear requirements and a solid understanding of the website from the start?
- Were goals and success metrics clearly defined?
Design:
- Were there sufficient user stories?
Resources:
- Were roles / responsibilities clear?
- Were there sufficient resources to execute all phases smoothly within the project scope and schedule?
Schedule:
- Was the schedule realistic?
- Did anything cause the schedule to slip?
Development:
- Was code freeze respected?
QA testing:
- Did development adhere to QA’s checklist?
IT / Deployment:
- Were there any issues getting the app onto stage?
- Were there any issues getting the app onto production?
- How did launch go?
Security
- Were there any issues in the security review process?
L10N
- Were there any issues in the localization process?
- Was string freeze respected?
Communication:
- Were the channels of communication between all the stakeholders established from the start?
- Was there anything that blocked or hindered communication?
Notes
Planning:
- FX4 campaign originally in October 2010
- Identified pain points, but couldn't change it too late in the game
- Product owner changed
- Cmore took over for ozten on July 5th, 2011.
Design
- Confirmation page changes late
- Object changes
Schedule
- Vendor had shared resources on their the project
- Vendor changed their lead developer three times
- Need strict start/end dates
Implementation
- 20 questions was too much
- Vendor was unable to duplicate our environment
Prioritization
- Webifyme was in flux
- Considered cancelling
L10N
- They were not involved or told about the project until 3-4 months into the project
- L10N resources were constrained with too many parallel projects
- The issues with L10N created a new external committee (World Ready) to get feedback on how Mozilla is doing with localizing.
- Spanish team lost all of their translations at one point
- Copy was too long and there were many changes after string freeze
- Be very clear on the exact locale for the translations (i.e. Spanish vs Spanish in other regions)
QA
- Team approach to QA
- First time using UTest for crowd sourcing QA work
- Difficult to QA when development is still blocking
Communications
- Internal and external people didn't understand why we were even doing the project.
Lessons learned:
- Get workflows in the contract
- Need vendor QA and proof before handing over any code
- Need a better way for a vendor to be able to simulate our infrastructure
- We can do external hosting, but only if it is a stand alone domain. moz.com|org domains must be hosted by us.
- Can we review code from an external vendor before a vendor is selected?
- When do we decide to outsource a translation? If it would ever happen, we need to do it in collaboration with the L10N community.
- Communicate the results of campaigns back to L10N community.
- L10N may use "simple English" as the root strings.
- How can we better communicate the goal of the campaign?
- We should do a brown bag for future campaigns to get more ideas and better communicate with people that want to know.
- sunk cost is not a good enough argument to drive a successful project.