Notable i18n changes: presuming it passes beta testing for the next week (done one already!) my split-domains work should make it in. Additionally, I’ve done a variety of things on my so-called “fixup path”: fixing edit type i18n, making the tabs on the statistics pages translatable,  and translating editor dropdowns. Non-i18n changes include bringing back voting icons on edit notes, fixing some statistics-related things (so, last year’s GSoC project), and fixing the database settings for slave installations.

Next week, during beta freeze, I’m hoping to start working on the language-picker and the associated reworking of our language-switching infrastructure.

Hopefully this’ll ship soon — but it’s already on the sandbox, for those who want to test it out.

Hopefully shipping the code changes for the new domains in the next few days!

I hope to update again shortly, once the translations are migrated to the new system (and once musicbrainz-server knows how to use these new domains).

Been a while since I’ve posted here, but for all three of you readers (±3) I figured I should do an update.

Specifically, I’d like to report that I’ve been accepted for a second year to work on MusicBrainz for a second year in a row as part of the Google Summer of Code program. Last year I worked on the MusicBrainz Timeline Graph (and statistics in general); this year I’ll be working on internationalization of musicbrainz-server.

MusicBrainz: Now in Japanese!

Internationalization is an interesting topic in this context – the basic structure for i18n has been in musicbrainz-server from the start, but it’s never been particularly well set-up, and it’s never run on production. I’d like to fix both of these problems. As you can read on the GSoC Project Details page for my project, I’m doing this through three distinct “paths”: interface, domain-splitting, and general fixups. Domain-splitting deserves further explanation: currently, all our translatable strings are in one huge context – more than 5,000 strings. I’ll be splitting them up somewhat to group things contextually related, both so some things can be reused (for example: names of relationship types, which are useful for anything using MusicBrainz data) and so that translation isn’t quite so insurmountable a task.

My first task, during the “Community Bonding Period” through May 21, will be to find “language liaisons” – people who can serve as people I can ask questions of for languages I don’t personally know, to highlight problems with our system for those languages, file tickets, and serve as coordinators of translation teams.

Read more on the GSoC Project Details page for the project! For the interested, our current i18n/l10n is available on Transifex. I’ll be updating that soon, though, hopefully. Some of our translations, a bit out of date, are also running on i18n.mbsandbox.org, an internationalization-oriented MusicBrainz sandbox (note: I also run the MusicBrainz sandbox server).

I hope to report more soon!

So I didn’t have much of a life at this point. In October, I read a post on a forum about NaNoWriMo, and went to check it out. I found out we had a local group and thought this was a great idea. After, all I liked to write! I always said so, except I never did it. So I went to the local kickoff and had a great time. I was 18, and would turn 19 shortly after the start of the month. There were a lot of people participating, but a core group had began to form. I was the baby of the group at the time.

I wrote 50,000 meandering words about a fairy named Kikichi, who is a character I still work with today. Aspects of that original NaNo tint the version of the story that now exists – there was a mute little human girl, and now Kikichi is the one who has lost her voice. Kavi, a winter fairy, was closer to the mute girl, and now he will be Kikichi’s close friend. We were dared to include “passing starbursts” in our novels, thanks to a candy bowl that was going around at the kickoff.

Of course I had to do it again the next year. I’d had so much fun. I had a laptop this time. I wrote the first draft of Circle of Many Faces, my greatest almost-novel to date, in my opinion – should I ever finish the current version. The NaNo draft is terrible and I will never show it to anyone, but even great things have humble beginnings.

In 2005, I was living in California with my new boyfriend. There wasn’t a local group and I lost interest after a few days. I wondered if I was done with NaNo.

I wasn’t. In my absence, quite a social group had come together in Tucson, and I found myself back there soon. I was having a lot of health problems, and couldn’t always make it to hang out with them, but I managed to do NaNo that year. I wrote a draft of another writing project I’d had trouble with, Searching for Sunset. It’s something I could definitely rewrite someday.

It was after that NaNo that writing regularly became a thing. I began to work on a children’s fantasy novel called The Rainbow Covey. That one isn’t done either, though it’s one that I have meant to finish. When I started having trouble with that, I started work on a new version of Circle of Many Faces. I stopped work on that for a month to do NaNo 2007. Ian (the co-author of this blog) and I got into a ridiculous race, and we each finished on November 7, close to midnight. That was my 23rd birthday. The breakneck pace at which we wrote that month (I wound up close to 80,000 words total, and had it not been for laptops breaking, I probably would have cracked 100,000) did not result in particularly good prose.

During the next year, I was on fire. It may have “helped” that my car was stolen just before Christmas, depriving me of transportation, and then shortly thereafter I went through a severe ankle injury, one that still haunts me today, and that deprived me of mobility. What else was there to do but write? I wrote nearly all of that book that year. I wasn’t quite finished by November, so I decided to jump ahead and start the sequel, The Murmuring Mirrors, for NaNo. I was very dissatisfied with both the quality of my writing and the plot that I came up with, save for a few scenes. I began to think that years of doing NaNo had perhaps trained me too well to write at breakneck pace.

I continued to write in 2009, though my pace was slowed, since I was back in school. November 2009 was filled with a horrifyingly busy class schedule and drama from a recent breakup. I began the month facing threats from my ex, and emotional abuse that – he admitted freely – was meant to sabotage my NaNo. I was writing Across, because I thought I could write a better specfic teen romance than Twilight. If I ever write Across, I think it could be. I was behind all month and had to pull 12,000 words out of my ass on November 30 in order to finish. Those 12,000 words are some of the worst I’ve ever written.

In 2010, there was no time. I got a job, but what had promised to be part-time turned into a 60-hour-per-week nightmare which granted me no real spare time to work on my Capstone, which was kind of like my senior thesis, for school. I was worrying that I wasn’t going to graduate. The reason I did graduate was that I worked myself into oblivion, letting a fever continue for three weeks instead of sleeping. Soon after, I was fired, because the bakery decided they didn’t need my help after all. I found myself wishing I hadn’t taken the job.

I hadn’t written much in 2010. School and life had been difficult and in the way. 2011 began on a much higher note; I was unemployed, but my new boyfriend was very emotionally supportive, and I took a few classes to fill my time. My art seemed like it was actually getting good, too. I began to wonder if juggling art and writing was bad, because I loved them both. I did start a story mid-2010, until school got too busy to work, and hadn’t touched it in awhile.

Mid-2011, I finally decided I wanted to write again. And now I write fairly regularly in small amounts. NaNo was fast approaching. While in the past I’d considered it a badge of honor to win NaNoWriMo, I asked myself, what was I proving? I can write a lot of crap really quickly, or I can slow down a bit, write all year, and write things that I like.

I’m not yet a success as a writer, but neither do I consider myself a failure. Maybe success is still lurking. Does it matter if it is? I don’t know. I know that without creativity I would be miserable, but I’m not miserable in my not-creative day job because I like the people I work with and because the job pays me enough to live, I’m not so desperate to find some way to make ends meet. I write less than I used to and draw more, because my drawings have value to me, too. I often illustrate the things I write.

Without NaNoWriMo, I might not have the discipline with which I temper my kind-of-chaotic creative urges. But I think the time for it is past, for me. I was only doing it to prove I could, but I already have proven I could, over and over. It’s time to move on. I don’t think of myself as having quit NaNo, I think of myself as having graduated. It brought me good things. Though the social group has by and large disintegrated, it left me with several cherished friends, including my best friend in the whole world, who is now my boyfriend. Without NaNo, we wouldn’t have met.

All good things must end, and I think I can feel good about moving on.

And anyway, I’m still writing.

So, hm, what all have I done:

a.) added some various statistics to the collection mechanism; specifically, artist countries, genders, and types. I made them get displayed, too! See http://musicbrainz.org/statistics and http://musicbrainz.org/statistics/artist-countries for these. I’m currently working on/have in review a fix for the gender calculation, since despite adding calculation for type statistics at the same time, I totally neglected to think of the fact that groups can’t have genders!

b.) I added copyable URLs (mostly) for the timeline graph. For example, see http://musicbrainz.org/statistics/timeline#-recording+-medium+-label+-artist+-release+-releasegroup for a graph of just Works. I have in review a fix for this one, too, so that the region you’ve zoomed to also goes into and gets parsed from the URL. In the same change, I made it so the zoom level doesn’t reset when you add and remove lines from the graph.

c.) In review, I have something to merge the checkboxes for adding/removing lines and the legend. Yeah, that means the checkboxes won’t move when you click on them any more! Part of this change was also hardcoding the colors (rather than hardcoding integer indexes into an array… complicated, but anyway), so there’s a new colorscheme which you can see at http://codereview.musicbrainz.org/r/1343/s/83/. Thanks to Eliza, my coauthor on this blog and fellow MusicBrainz GSoC student, for helping me work around my stunted color sense.

d.) In the works:
i.) MusicBrainz Events, once I figure out how to display and interact with them in a nice way.
ii.) Statistics for release and label countries (in the same vein as artist countries), and potentially a whole bunch of other things like instruments, formats, scripts, languages… etc. This is a bit removed from my actual project, so the progress will be a bit slower!

So, cool. Now to see if that run of CollectStats.pl has finished up yet…

Or, if you don’t want to go over to Review Board for yourself:

This thing’s getting close to something that might actually be useful!

Now I just get to wait for my mentor to review and merge my code. Wonder if there are any low-hanging bugs I can tackle in the mean time…

Right now it’s a pretty basic graph (only Artist/Label/Release/Release Group/Medium/Work are shown, and the controls are a bit primitive, and a bit undocumented), but it does exist, and lots of improvement is happening. Since the part that’s in master, I’ve already added better controls, better axis labeling, and some interactivity on hover.

But hey, shipping code 8 days into GSoC — I’ll take it!

For those of you new to this (even though it’s only the second one!), this here’s a little series on what I did this summer. As I disclaimed previously, not everything here is necessarily a summer-only project; in fact, after this particular installation, I’ll be moving along to things that I started during the summer, but which I haven’t really done too much work on yet.

This particular project started in February as a class, continued through the summer as something simulating a “real job”, and will be continuing through this fall as a work-study position.

This one’s about TheHub, my school’s own special academic webapp, and the project I’m part of rewriting it.

What is it?
TheHub is the online academic software for my college, Hampshire College. Previously, we’d been using the packaged webapp that came with Datatel plus a lot of custom code (that code being necessary because of Hampshire’s rather unique academic program, which you could read about elsewhere). It was quite a mess as code goes, having been cobbled together over many years, and an interface disaster to boot. The administration finally noticed and decided that the way to fix it wasn’t the expensive option of “get someone to write us a new webapp” but rather to talk to some staff and faculty members who had been thinking of how to improve it for some time – and their way of fixing it was to start out with a class.

Tell me about this class…
So, the class. Last spring we had a class, which was mostly consumed with designing, both interaction-wise and visually, how TheHub was supposed to work. There was a small subset of us, myself included, doing more in the way of technical infrastructure instead, but the main thrust of the class was to bring a bunch of people together to think hard about what the software should look like. We used Alan Cooper’s methodology of interaction design, at least partially, and worked through what we needed.

It turned out okay, at least. There were useful results from the class, which is more than many of us were expecting. One of the most interesting results is our visual design, in fact, which is now being adopted by other parts of the school (specifically, our Moodle install and our webmail).

So what then? Wasn’t this a summer project?
Funny you ask – yes indeed! Three of us from the class got hired for the summer to keep working on it; specifically to work on actually implementing a lot of what we’d designed. It was at a rate substantially lower than most contract programming jobs of course (that being part of the appeal, for the school), but of course we aren’t exactly veteran engineers.

Fairly straightforwardly, we worked on implementing that which had been designed. That which hadn’t been sufficiently designed, we designed more. It didn’t make it perhaps as far as any of us would have liked, but it got far enough that there was money provided for the fall, where we’re continuing to work on it. Not much to say there, I guess – we’re churning away, however fast (or not).

So, uh, technical details?
Of course! We’re writing it in python using the web.py framework. We chose python because it’s a language that’s being taught to a lot of people (so once our current crop leaves, finding maintenance programmers among students will be possible), and because it’s generally a flexible language (so that when our academic program changes, again, in two years, it won’t be too hard to shift around our software). We chose web.py because due to things beyond our control (such as DBMS) we would lose out on a lot of the features of a more-traditional framework; with web.py we can piece together something workable despite sometimes-strange constraints. We use subversion for version control; this is because during the class we wanted it to be highly accessible. I’d have preferred git (my personal favorite) or another DVCS. Oh well. It works, mostly.

Where is it now?
Still working on it. I’ve been having a busy semester and haven’t personally contributed as much as I’d like; the same has been true of the other student workers – but it progresses nonetheless. Our code is sometimes nasty, but that’s improving. We haven’t replaced all the functionality we need to replace, but we’re getting there.

In short, I’ll have to tell you in another little while how it’s ended up. Stay tuned?

So. There’s TheHub; hopefully my explanation, like the software itself, is not too bogged-down or enterprisey!