0.5 Retrospective
About the Retrospective
The Fluid 0.5 retrospective meeting is an opportunity for community members to get together and talk about what worked well and what needs to be improved in our shared design/development process. Based on this reflection, we'll come up with specific actions and we'll volunteer to be champions for specific tasks. A champion is not necessarily expected to accomplish the action, but rather to help remind people of the importance of the goal and ensure that it stays on our radar.
Agenda
Look back on the actions and results from the last retrospective.
Look back at the highlights from the past two months.
Collaboratively answer these questions with a backwards looking perspective:
What went well?
What didn't go well?
What did we learn?
What still puzzles us?
Collaboratively come up with specific, concrete actions
Actions we implemented
More retrospectives
Design team on IRC
More paired designing
Put things we've learned into practice
Help people understand our processes
Move the standup meeting to a later time
Polish up our research findings
Actions we partially implemented
Whiteboarding tool
Automated testing
Get components into other apps
Bring in more a11y into design process
Share out personas
Revisit this page
Templates for component design pages
Actions we didn't implement
Get other people creating components
Try out Balsamiq for prototyping
How to contribute patterns
Find a better way to release tech docs
Create a good Fluid primer: tech and more
Design highlights from 0.5
Wiki
Tons of improvements to the wiki
Repeatable, easy to navigate, structure to component pages (ex.Inline Edit Family of Components)
New component icons for web presence
Improvements to Fluid Website
Research Models
Personas released
Frequency added to "Use Case Frequency Matrix"
Components
Renaming Reorderer and Inline Edit families
Began work on component keyboard interaction summary table
Simple Text Inline Edit in Section Info Demo
Began working with Michigan to integrate Reorderer and Date Picker in several places in Sakai
List Reorderer design started
Date Picker designs making crazy progress
Uploader design enhancements based on user testing and revisiting the design framework
Layout Reorderer user testing and performance improvements based on results
OSDPL
Architecture improvements continue to be made on the OSDPL
Accessibility improvements based on review
Governance final
Began defining "How to Create a Design Pattern"
Other
Shared out and began using design framework (continual improvements)
Skype team chat
Development highlights from 0.5
Framework
DOM binder implemented
events system in place
skinning system started
Reorderer
API's changes and migration tutorials
replaced jQuery ui.droppable which led to speed and consistency improvements
Safari support improvements
integration with uPortal
Uploader
uploader 2 work begun
Other
functional examples
0.5 beta release
0.5 release
jQuery accessibility started
What went well?
Framework: it's here! And more importantly, it's not framework for framework's sake; our components evolved alongside it, and their needs drove its design
The Infusion 0.6 roadmap was created early: well before we started
Renaming and reconsidering the component families was effective
Implemented the design framework tasks and actually used them: nice to have a big picture in mind
Date picker designs started
Designer Skype chat is really useful
Wiki reorganizing is nice, pages are looking a lot cleaner
Improvements to the Reorderer are great and have led to improved integration with Sakai and uPortal
Communication on list and on the chats have been really incredible
Great to have the VULab integrated into the community: having Blake help with OSDPL, for example
Release process is really fun and effective: fast, energetic bug fixing
Seeing components get integrated into Sakai
Collaboration with Michigan was really effective and thoughtful
Lots of "just do it" tasks accomplished: increasing confidence in our work and how it fits into other tasks
Getting more comfortable expressing ourselves visually: eg. progress indicators, component families, framework
Lots of positive feedback about Infusion 0.5
Perhaps we bit off more than we could chew, but this wasn't necessarily a bad thing: we pushed through with a very solid release
Justin's contribution with QA and community coordination: he's quickly catching issues as they arose, and shared them out
What didn't go well?
Would have been better if PreferAble had been in the design cycle for a little longer.
Documentation and communication about the framework could have been better: we should have taken the time to write and talk more about the framework as we coded it
Scoping: being realistic about what we can do in a particular release
Bug wrangling near the end was a real challenge: too much scope for the release?
Restructuring the wiki was a challenge: content got buried accidentally
Scheduling is always a challenge
Automated acceptance testing has been a real struggle: a lot slower than we would have hoped for
Communication: signal/noise ratio
Framework was moving too quickly without documentation: was hard to play catch up
Planning is really good, but we can still improve on it: design and development more in sync
Long-term planning in JIRA can be time-consuming
Still not doing enough quick visualization of our work, community, code, and designs
There's still room to improve our communication: this will be a continual problem, but is a good thing
Left Reorderer tests failing for too long
What did we learn?
The jQuery UI community is a fun place to play
How to collaborate with partners; how communities are managed; details about our framework as it emerged
Talking with other people and other open source communities
how the design process fits together
some of our work happens in sprints and some are distance runs
some of the early activities in design are very important.
speak up about frustrations early and come up with solutions
trying to do too much work at one time causes chaos and rubble
Reminded once again how nice it is to build UIs in a dynamically typed language: views, events, and callbacks just fit
Learned about co- and contra- variance in OOP: interesting, but not that interesting
What still puzzles us?
How people are going to pick up our stuff and use it in the real world
How we have component leads that aren't bottlenecks: how to share our work better
How to satisfy multiple audiences for the wiki and OSDPL
How to implement testing strategies, and getting people involved
Where is collection space going?
How do we satisfy different audiences?
How do we make our sprints less painful and less frequent?
How do we make our work more of a distance run that culminates in a release?
How do we make design work for components that live in different problem spaces? Personas are an example of the difficulties.
The event framework is puzzling
How do we get all of 0.6 done with the distraction of a face to face
What is the future of Fluid?
What does the full extent of our declarative approach look like for setting up components?
Actions
Action | Champion |
|---|---|
Come up with a documentation strategy: how to keep it up to date, archiving, etc. |
|
Bring up issues as soon as they come up: communicate about problems or messiness early |
|
Create ways to ensure the design team is more in touch with bugs and issues |
|
Code and framework "heads up" emails out to the list early and often. Echo conversations back out to the list. |
|
Help others build Fluid components |
|
Documentation: How to contribute patterns? |
|
Automated UI testing |
|