5
edits
Changes
Summer of Code/2015/git backend alex (view source)
Revision as of 00:58, 23 March 2015
, 00:58, 23 March 2015Created page with " = WORK IN PROGRESS = = About You = Name: Alex Herrmann<br /> Email: alexandermherrmann@gmail.com<br /> Wiki username: alexmherrmann (Alex Herrmann)<br /> IRC nickname: ale..."
= WORK IN PROGRESS =
= About You =
Name: Alex Herrmann<br />
Email: alexandermherrmann@gmail.com<br />
Wiki username: alexmherrmann (Alex Herrmann)<br />
IRC nickname: alexmherrmann (freshly registered)<br />
First Language: English<br />
Location: Salt Lake City, Utah<br />
Time: Weekends I am always free from 10 AM MST (4 pm UTC) to 10 PM MST (4 am UTC)<br />
=== open source experience ===
My experience with open source has typically been that of a consumer, Most of the open source projects I have were small pet projects
from early high school, before things got busy. Now that I'm attending college (at the University of Utah) my interest in open source has been re-ignited,
I have a little bit more time and freedom to code as I please and contributing to a project like sugar would be extremely cool (Our family actually has one of the original XO laptops purchased when they were first made available, we moved a little while ago, it's probably still out in our garage somewhere). I chose XO as one of my organizations because beyond being able to productively code for a summer, I have experience with the XO laptops,
= My project Proposal =
Having a strong background in git (my old testbed where all of my code went is at github.com/alexhairyman/ unfortunately it has not been contributed to in a long time as high school and a job took over my free time), I would love to implement the Git backend for sugar journal. Essentially I would want to modify sugar journal's code to essentially propagate all changes as a commit to a local repository. The end goal would be to wrap up git with a UI to explore all previous changes, each commit would contain a metadata file generated by either the datastore or the journal source code, and then perhaps a binary file containing the content too. This would update a locally stored git repository, which would then propagate changes to any one of the git hosting sites out there (github has a nice python API to facilitate using github specific features). One way of doing it could be store all of the binary/large data to google drive, the API available for drive would make it easy to store data in an existing drive setup. Instead of using git at all, having a metadata file keep track of the journal entries would allow the computer to just interact with google, and expanding storage for google drive is ridiculously cheap. This would be the first discussion I have with my mentor
= community =
'''What will you do if you get stuck and your mentor isn't around:''' The nature of my project will allow me to leave a message on either IRC or as an email for mentor, and while they're busy, there are enough facets of my project so that
=== Rough Timeline ===
* The first few weeks (no later than may 15th) will be when I explore, in-depth, the journal and datastore and talk with my mentor on the best possible way of "gitifying" the events and journal entries
* the next few weeks (mid May to beginning of June) will be modifying these to produce an output metadata file each time a journal entry is created that is associated with some kind of binary or compressed content. This will be either pushed to git or google drive. I'm thinking a small custom format that is human readable is good, I'm thinking YAML could be especially good.
* the beginning of June I would hope to start stress testing the system as no more than a local repository
* half way through June I would begin the remote propagation part, where content hosts such as github, google drive, imgur would come into play. This is where the meat will be, keeping track of what's been pushed and how to access it. I am aiming for google drive first. Getting hooked into the API with python to store the binary data, and then having one large metadata file or do it the git way
* This will take a while, interfacing smoothly with the API's of the other services. I estimate up to late July
* The final weeks of GSOC in August will be the testing and integration phase, where hopefully community members can get their hands on it and test it out, see how well it fairs in the "real world"