User:Mchua/Braindumps/Learning plan

What do I need to learn in order to become a good generalist OLPC hacker? (More importantly; what do I need to make to learn what I need in order to become a good generalist OLPC hacker?) Note that this would by no means make me an expert on anything, but should give me at least a decent idea of what people are talking about. Also note that although this is a "6 month plan," it would take far more than 6-months of full time work/study to get through it all, so it's really several years of hobby-hacking time given my current non-hacking OLPC grassroots volunteer workload and the impending need to acquire some means of purchasing food.

THIS IS A DRAFT. I don't know what I'm talking about; it shows. Comments are extremely welcome. Feedback even more so. Constructive criticism most of all.

Device driver

• Write a device driver. (What for?)
• This will be... painful.

Linux

• Grok it better. (Yeah, this is vague.)
• A mildly ambitious plan might be to contribute to a kernel patch, but this is by no means the only way of going about it. And honestly, this would probably take me way more than 6 months. My knowledge of kernals and OS design is nil, and my C horribly rusty.
• Start with a re-tour through K&R to un-rustificate my C.

X / Display

• I don't even know where I'd start with learning about this - I don't know what has to be done.

Packaging

• Learn how to make a package - probably a Fedora one.
• Trying to package Sugar (as Jani did with the Ubuntu package) might be a good and useful, if somewhat challenging, start.

Applications

• I interpret this to mean Activities.
• Write (another) one (that's actually useful, not a silly game where a bird poops on cars.) Use Pygame for sanity.
• Implement mesh networking on it...
• ...and then make another one in pyGTK, and add meshing.
• This would probably take 6 months on its own, if not more.

GTK

• Get to this through building a pyGTK activity; see above. Then decide where to go.

Mesh

• I get the feeling that this would involve going back through Analog and Digital Communications 101 before anything made sense.
• Look at this more and find a better place to start.

Power management and charging

• Don't know where to start. Should look into this more. What does OHM do? What does it need?
• One project I should do, though, is "build a charger."

QA

• I would probably start out by getting a team of volunteers to work with me on exploratory testing for the current software - exploratory testing is what I spent most of my TOPP internship doing.
  • It's fantastic for getting good bugs out of the woodwork in software that had to be released yesterday (and is rapidly changing tomorrow).
  • ...without crushing the souls of your testers. (Really. I hated testing until Tim Coulter taught me this method.)
  • I'd seek constant guidance and feedback from the Much More Experienced Testers in the exploratory testing communities that I've started to get to know.
• While doing this I'd look for ways to learn about other aspects of QA; there are many. And then flesh out this plan-segment more.
• On a fun mechanical-engineering foray, I still remember a few things about testing mechanical and material properties of products from matsci lab.
  • Most of them involve breaking stuff.
  • Slowly and scientifically, so it's not as fun as it might sound, although impact testing is plenty awesome.
  • I've done this twice before as matsci projects; once on bicycle frames, once on a CD drive.

Performance analysis

• Still unclear on what this means or what's needed.

Community tools

• Make a useful webapp.
  • Like the actual Bounties for Awesome site or something like it.