Talk:Summer of Content

From OLPC
Revision as of 10:52, 24 July 2007 by Php5 (talk | contribs) (Availability of tools)
Jump to: navigation, search

What this page is: This is a place to discuss the current Summer of Content 2007 session of Summer of Content as well as the Summer of Content program overall and what we could do with it in the future. Post your ideas, questions, and thoughts! Click here to leave a note.

You may also be looking for: For discussions on the northern and southern summer sessions this year, see #Summer of Content 2007. For older discussions (warning: lots of old and outdated stuff!) see the archive.

How to sign up to mentor/be an intern

Where can I go to join the program or submit an idea?

To submit a project idea, check out the Summer of Content project ideas page. We're building the sign-up sections for interns and mentors right now, and they should be done by the end of the day (Wednesday, July 18). If you want to be notified when they're ready, leave your email contact information at the bottom of this page and we'll send you a message when they're ready to go. Mchua 10:04, 18 July 2007 (EDT)

Summer of Content 2007

There are two 2007 sessions planned - an initial pilot season over August and September, and a session starting in December and running through February.

Contact me when the applications are up!

notes on orthogonality

projects have multiple authors
multiple interested interns
multiple experienced/interested mentors
interns have multiple project interests
interns have one personal application and description of interests
mentors have one personal applicatoin and description of experience
there is no restriction on projects <--> applications.  
 a project may end up being proposed by an anon and never have any interested mentors or interns

on nesting templates

Check out the design of the HIG. Eben and I spent a while working out the right recursively nestable page structure so that each page makes sense on its own, and has sensible headers, some of which show up on the leaf page and others which show up in transclusion. if you want to include header navs, make sure they are noincluded. Sj talk

Need focus on joining the community

Karsten Wade, Red Hat (paraphrased): One thing we ran into with the Summer of Code was a lack of Collaboration. Projects need to be something that works with the existing community of the open-source organization. It's hard for mentors when students go off in a corner and work on everything themselves instead of becoming part of the community (it doesn't create "less work" for the existing developers, they want new people to take the time to join the community, not just crank out a deliverable that might not work for the existing culture.) Can fix this somewhat by giving people parts of collaborations instead of selected individual projects, make it clear that it's their job to partner with other human beings.

We've tried to do this with the design of SoCon - need to make sure that community-joining is explicitly in the job description.

Meta-projects

Another suggestion from Karsten: Summer of Code recruits individual interns, why not recruit teams or clusters instead? Have one project be in charge of communication between other projects, one in charge of reuse and accessibility of other projects... make "meta-projects" that facilitate usage and community building for other interns.

(now you know where the meta-projects idea comes from!)
What is the right ratio of meta-projects to other projects? 2:5, 2:10?

Bounties

Another solution for matching content problems and contributors are bounties. They already exist for open source problems, originated with mathematics problems (solve X theorem, get $Y - both large and very small amounts for problems ranging from Fermat's Last Theorem to "prove this tiny easy lemma"). Have people post bounties with dollar amounts; start a distributed system.

  • Josh Gay: I think bounties do really well in India; many are picked up there. They've been displaced a bit by low contracting feeds by companies that have been picking up small tasks.
  • Can do this around Summer of Content to jumpstart development - maybe forming bounties around proposals that are not selected from this round.

Trustworthiness?

  • What about trustworthiness, etc. how do people guarantee goodwill and that things get done on both sides of the transaction?
    • Have a company act as a clearinghouse for bounties? (problem of maintenance, high overhead)
    • "Just trust people, things will work out" - what if there's goodwill on both sides but people disagree on what it means for a bounty to be "finished"?
    • Mel Chua and Andy Pethan working on a potential solution, see Bounties for awesome project

Problem - lack of community?

  • In open source, this doesn't have a good rep. Same "community" problem discussed above; they're one-off jobs, someone pops in to do something quick, doesn't always take the time to integrate into the development community, get to know the culture and people they are working with.
    • Huh, I thought this was a plus of bounties and of Summer of Code - it's like handing someone an API to your community and saying "here, as long as it interfaces in such and such a way it's okay" and not having to spend a lot of time bringing a newbie in the long way. I guess I was wrong. Mchua
    • SoC gives multiple months coming into the summer to integrate. If the "join the community" aspect of the job is made explicit, then it makes sense to have bounty levels, or a brokerage, since the people there will already be invested in the community.

Celebrity-sponsored bounties

From Josh Gay: Another interesting thing: in math, it's not just the problem that is interesting, it is the person who poses the problem (and isn't able to solve it). For instance, on the Summer of Code forums, there's just as much discussion around t-shirts and certificates as there is around code.

Branding does count. I met some developers who did Summer of Code last year and I asked them why they did it; their first reaction was "because Google is cool, and I wanted to be part of something that Google was doing... because they're Google." Name brand recognition! Mchua 18:27, 22 July 2007 (EDT)

Josh Gay: This can be an opportunity for us, since we have a large mindshare, to reach out to sponsors: famous people... not to necessarily contribute time or energy, but to put their name to a set of bounties that they care about. kids don't just get 10 bucks, but get that from a local soccer player with a name attached to it. this is a good opportunity to set this up properly for people always going around, like nicholas, to say "hey! if you want to help, put your name to these and kids get a thing thanking them from you.

Example potential bounty sponsors: Shuttleworth, Shakira, Bono, etc. put $10 down (or some other tiny sum) on a project they care about.

Karsten Wade: Bragging rights are important, have a little flag or something you can post on a website that can show how many problems you've solved (bounties you've claimed). the trouble is the only way to really be successful is if it's done in a truly open way (people can see the things you've done to warrant "bounty flags") - so you can make a bounty that's translating the subtitles of something into another lang, or editing the braille output of a particular doc or doing those particular things... which aren't necessarily that glamorous.. bounties might help by attaching meaning to them at the same time.


It's hard to struggle for individual student slots as a mentor -- so that other projects get lopped off.

Tools

Some possible tools from teleconference brainstorming:

  • bugzilla wedged into a review/proj mgmt tool...
  • xml/docbook...
  • content-related scripts that anyone can use
    • central repository for version #s and snapshots.
    • making snapshots / defining repository structure.
    • consider maintaining content as svn trees. time and date?
  • Documentation about how the software-build system works.
    • devs or others might be interested
      This requires a decent amount of technical knowledge even if the end-product isn't "technical." Would be a great project for an engineering student who wants to stretch out a bit beyond "traditional" technology work, we need more of those. Can we reach out to software, hardware, etc. engineers, scientists, mathematicians who want to do things other than engineering/math/science, as well as arts/humanities/social scientists who want to get more involved in technology? Mchua 18:27, 22 July 2007 (EDT)
      Remember, academia isn't the entire world. There are people around who aren't planning on being professors when they grow up...

Availability of tools

Karsten Wade: Bounty systems (and open content work in general) wouldn't work if everyone had to have a copy of a toolsuite and a computer.

From teleconference: Tools need to be easily available, accessible... we're still biased towards people who want to spend all day in front of a computer (and who have a computer in the first place).

This is where outreach comes in; have some of the computer-people take on a project to reach out to people without computers - not to get them to use computers, but to act as a bridge between what they're doing in the physical world, and what's going on in the internet world. For instance, someone who will take pictures of the work of local artisans, and make a gallery of them online, or film local musicians and work with them to put their clips online... help people join the worldwide network without having to spend a lot of time in front of a computer. Mchua 18:27, 22 July 2007 (EDT)

Suggestion

The Summer of Content page had better have {OLPC} template on it to avoid any misunderstanding regarding to the scope of participation. php5 10:52, 24 July 2007 (EDT)