- 1 Automating community statistics: datanommer
- 2 Improving Fedora's brand guidelines
- 3 commarch.usersys administration
- 4 Fedora Project Wiki localization setup
- 5 Making POSSE's application process not suck
- 6 Keeping up with Fedora
- 7 Other noteworthy things (read: everything else I can remember)
I work for Red Hat's Community Architecture team as a general programmer and guy who does stuff out in the community.
— Ian Weller
Aug 04 2010
Automating community statistics:
At the end of the summer, I have a much clearer goal of how I want to (help) automate the gathering and analysis of statistics within Fedora than I did at this time last year or even at the beginning of this summer.
Instead of attempting to solve every person's statistical questions in a web interface that does not update very often, we're writing code that grabs information from each application or "thing" we use in our community to communicate and get things done. A subset of that information will be generated into graphs automatically on Fedora's infrastructure, but the bulk of it is meant to be run locally for one's own research and hypotheses.
The application is called
datanommer and is community-neutral — currently the only data-grabbing code is for MediaWiki-powered wikis, and it's been tested with fedoraproject.org/wiki and the English Wikipedia (works just fine). It's written in Python, which makes it the easiest for others to contribute to.
The next steps:
- Write more code to download more data from more applications
- Write more code to create more types of graphs
- Write documentation
- ... especially on creating complex queries (chaining different pieces of data together)
- Fix the ugly code in cli.py:main
- Get some data analysis integrated into Fedora Infrastructure
Improving Fedora's brand guidelinesI spent a lot of time between last year's internship and this one tracking down original SVGs for all of Fedora's common logos, cleaning up the SVGs, and writing a new revision of the current logo usage guidelines as new brand guidelines, encompassing logos, colors, fonts, and more. This has been done in cooperation with the Design Team and Red Hat Legal (props to Pam Chestek).
There was an attempt this summer to get this done, but I had so many things that I was working on at once that this item got neglected a little bit. What's done:
- All logos with a TM symbol in them (Fedora logos and Fedora Remix logos) have had their SVGs finalized.
- Changes have been approved to adjust fonts in the FUDCon logo and Foundations artwork.
- MgOpen Modata has been deprecated in favor of Comfortaa and Cantarell (with a fallback to the Droid Sans family for Cantarell).
What needs to be finished up:
- Body text for the guidelines — just having pictures doesn't make it obvious enough to most people what you mean.
- Changes made to adjust fonts in the FUDCon logo and Foundations artwork, and approved by the Design Team folk.
- fedora-logos-free.tar.gz tarball made and hosted somewhere.
- Guidelines moved out of draft and announced everywhere and anywhere. Add pizazz about doing brand the open source way.
Máirín Duffy hasn't had the time to flush this queue as often as she liked, so she gave it to me. It's been slower than I expected (mostly spam). I answer questions and hand out fedora-logos.tar.gz.
Even though we do most of our work in the community, CommArch runs an internal server (an old machine that sits under a desk) that makes our lives easier. The server was last on F11 and needed a bit of love. I took over the administration of the machine and have been (slowly) documenting it on the machine's wiki.
(I spent a good half hour writing crappy PHP/HTML to create a "vuvuzela mode". If you're a Red Hat associate, I implore you to try it.)
Fedora Project Wiki localization setup
Spent a day getting pissed at how localization presently didn't work for the wiki again and decided to drop everything and fix it. Ported some templates from mediawiki.org (same license as our wiki, no worries) and announced the setup later that day. It'll be interesting to write something for datanommer to find how many translated pages we have using this new system. It's not a perfect system by any means, but the most important thing is that it works, it's intuitive (mostly), and that it works. (Also, it works.)
Making POSSE's application process not suck
I'm writing a Django application to help Mel be able to offload running a POSSE onto other people and make her life easier. The 0.8 release is slated for the end of this week (Aug 7 2010ish). It allows people to register for a POSSE, allows people to run a POSSE and approve registrations, and allows Mel (and friends) to approve an entire POSSE. Sends confirmation emails. Works a lot better than an email queue. Awesome all around.
Keeping up with Fedora
- Packaging: most packages stayed updated, others eventually need orphaned. Completed a few package reviews and got a new packager sponsored.
- Wiki czaring: generally helped people out, did the localization stuff (listed above)
- Design team: helped Máirín out with meetings she couldn't lead
- Websites: random commits to fedora-web.git to help people out, most recently the "not getting notified about software updates" global message
Other noteworthy things (read: everything else I can remember)
- Went to SouthEast Linux Fest, talked about datanommer, changed how I was going to write datanommer a week later
- Went to the Community Leadership Summit in Portland with the team and was thoroughly disappointed
- Inflamed my addiction to barbecue and sweet tea
- Got sick my last day at the office (bad Burger King for lunch, ugh)