[Greenhouse] Development schedule

[David Lu]

This is the current development schedule. Version 1 are the goals at the
midterm evaluation, version 2 are the goals at the final evaluation.

Version 1
-The first version will be focused on the Debian community. My first
priority is to get the code for Ubuntu DAT (Developer Advisory Team)
working, get it into Git, turn the dependencies into a req.txt file, and
get the UDD (Ultimate Debian Database) dump into PostgreSQL on my computer.
-The second step will be to write out the research problem statement after
the port and talking to makohill, and then assembling paultag and other
mentors to use it.
-The third step will be to create a better frontend. The current
inspiration is Trello. The "columns" will represent experience from least
to most. Each contributor will be represented by a "card". Hovering over a
"column" will display metadata about the "column" and hovering over a
"card" will display a preview of the contributor. Clicking on a "card" will
bring up a dashboard of the contributor with an events feed. Other features
to add include promoting contributors and keeping notes on them.

Notes:
-we decided the name of the project will be Greenhouse, and the tagline
will be "Track your project's contributors in the greenhouse"
-UDD dump will be 8GB on PostgreSQL
-DAT assumes people are only worth tracking when "sponsored"
-DAT gets fresh data once per day
-add an ignore_reason column for contributors with a foreign key to reason
codes
-use the ratio bugs:packages as an initial metric for retention rate
-possibly use deb_tags to recommend other packages once a new contributor
finishes a task
-create a Python script that fetches UDD data and serves it as JSON from
the web so only needs one adapter

Version 2
-Create an adapter for importing data sources and writing scraping scripts.
The first scraping scripts will be targeted at the Wikimedia Foundation.
Write documentation for plugins.
-Improve the design by diving deeper into Debian sources (irc, bug filer,
email list, screenshot list, translator list).
-Unify identities for people who participate with multiple identities and
add access control.
-Talk with makohill and improve data models with behavioral studies. Divide
into experimental and control groups to test retention strategies.

Notes
- should the adapter be a separate app with a data_pipeline setting?
-other interesting things to check out and possibly pull data from include
popcon (popularity contest) and lintian (code quality checker)
-all Ubuntu contributors have a Launchpad ID while only Debian Developers
have an ldap_id, so most Debian contributors can only be identified by
email (and many have 2 emails)

Extras
-If time permits I will port the oh-mainline code from Django 1.3 to Django
1.5. Django 1.5 allows custom user profiles, so this will be especially
useful for the Profile code. The bug tracker should be imported.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhatch.org/pipermail/greenhouse/attachments/20130621/24c2b986/attachment.html>