The Future of Collaboration - Daniel Greenfeld¶
Note
Audreyr took these notes. :)
Intro¶
Danny cartwheels, still blogs, works on Django Packages and whitespacejobs.org
Mark Pilgrim is gone
- feedparser, httplib2, Dive into Python, Dive into HTML5
- How much did we lose with Mark leaving the developer community?
- kennethreitz created a mirror at https://github.com/diveintomark
Where is httplib2?
- PyPI? No
- Not Google code
- Hard to find cached download
- Many libraries depend on it
Repeating history?
- django-piston, python.org, opencomparison.org all have bus factor and need active maintenance
Dark future?¶
- Critical Python packages vanish
- Build scripts fail
- Can’t always replace from caches/backups
- Legacy projects unmaintainable
- Domain knowledge leaves
- Hard to move forward
- 3rd party community as critical as core
- Actually, this is not the future. It’s today
Like the Library of Alexandria
- When we lose our history, we lose ourselves
Trust issues¶
- External and internal social issues
- Makes collaboration hard
- Causes “Not Invented Here” plague
Solutions?¶
Sponsorships
- But focused on short-term development, unusable code from interns
- Server costs are not the issue
Community managers
- Needs core/senior devs
- They’re busy
- Volunteers have different priorities
Paid community managers
- Work with package authors/maintainers
- Mitigate social issues
- Precedence: Ubuntu, Fedora, Twilio, Github
How do we keep Python’s community projects active?
- PSF project incubation: YC-style seed funding
- Helping market projects via python.org, blogs, other channels
- Help community projects find a business model, sustain themselves
- Copy startup model for projects that benefit Python