infryq

Yesterday I helped diagnose and fix some problems introduced by the massive production merges from Thursday, went to the all-hands, then had one-on-ones weirdly spaced all afternoon. The UNION approach turns out to have too much latency to be usable, so we’re preparing to merge the big and little tables next week. I had C shadow me building the document for reviewing February goals and planning March ones, then met with J to finish building datasets, which took ages but I feel good about the results.

Dinner, crossword, SG-1, snack, sleep.

Hell’s Rejects continues bizarrely, turning into legit religious sf based on Christian mythology. It is a weird mix of slapstick irreverence and deadly seriousness.

Yesterday we learned that the improvements we pushed to metadata resulted in a 75% reduction in running time, which is amazing news. We now have enough runway to make all the improvements we want to how the covidcast data is stored. The rest of the day was meetings, notably including a quick orientation for C on what lab manager duties might be like in future. I also worked with A and B to do a database migration for the new version of the hhs hosp data schema.

In theory I could have gone to dinner at 6:15, but instead brain decided to clear up some work details that really could have waited until morning, and I didn’t actually quit until quarter to seven. Dinner, trash, crossword, snuck in an SG-1 episode under the wire, direct to sleep.

Dreamt of guitars, tattered fiber processing equipment (thanks subconscious, I’m aware), messy rooms, running workshops, turning stitch markers into earrings.

Ambassador 1 continues; the author is a much better linguist than biologist.

Yesterday I worked with K to reorganize the way we handle data licensing in the API documentation, so that adding a CC BY-NC license to the claims signals as an exception would make sense. I went to the normal meeting with the claims folks, met with Engineering, and with the high school robotics girls, and in the cracks and afterwards finished all the work to bring the claims signal back online. I also had a working meeting with S to figure out how we want to organize the Asana-GitHub sync.

Late dinner, and I got overly chilled and was just fine with the day. P was lovely and handled the trash, then we did the crossword and I watched a little Avatar and went direct to sleep. Dreamt of a cat park — like a dog park, but all cats who are leash-trained. Pretty excellent.

I finished City of Illusions; it was okay. An abrupt but appropriate ending. Some unfortunate language implying deafblind people are incompetent.

!markdown

Yesterday I met with team leads and discussed the current state of hiring and next steps, met with new member D about the combination signal, met with Engineering, found a new bug in the new F codebase, then a bunch of shit hit the fan in the doctor visits signal and I wound up working until 7:30+ to help resolve it. I wrote a thing to manually adjust the metadata for doctor-visits back to 3 August, B helped test and deploy, and once that was in and the map was properly convinced the bad data didn’t exist I dropped it from the database. I helped our comms team develop an accurate message to send out to the list and to everyone who wrote in about the problem, then called it good. There’s still to figure out who will run the pipeline once good data becomes available, since M is on vacation, but that could wait until morning.

Snack for dinner, crossword, sleep. Dreamt of preparing a massive crab for a special brunch, two days worth of dirty dishes, tricking mean girls with exploding eggshells stuffed with confetti and chili pepper, dissolving go stones, and an infinite supply of Tupperware nut choppers, no two exactly alike.

Yesterday my laptop again refused to ssh to campus. The spare Air also refused. P’s machine agreed, so I borrowed it for the twenty minutes it takes to look over the aggregation results and commit them. I had a morning full of meetings, so I didn’t need it until after lunch, when it apparently also started to refuse to ssh out. P rebooted the whole house network, and magically everything worked again! Hooray. I met with L, a new volunteer with a process improvement background; we’re having him do some basic monitoring of the map site for now but will probably put him on some more technical tasks next week. I went to therapy, which was good; it helps me sort out what things are genuinely wrong versus minor annoyances, and reminds me that I am in general a ridiculously agreeable person mysteriously robust to the worst of the downsides of that trait. I met with F and the team from M about how to handle the difference between recorded time and start time, and give an overview of the smoothing code. I met with the indicators team and we had a long discussion about only having bad options for computing standard error for particularly small regions on one of the particularly odd metrics that makes geographical pooling difficult. I went to the team leads meeting and gave our update. I then got stuck in an analysis problem and it was suddenly 7pm, so I wrote it up, sent it out, and called it a day.

Dinner, crossword, direct to sleep. Slept ridiculously poorly; combination of bad body heat management and surprise period cramps. No one’s favorite thing.

Yesterday was bananas. The morning started out alright; I got a draft finished of a janky way to get region-specific covid benchmarks up. The lab meeting was a little unhinged; the PIs were half an hour late from a meeting with the university president that ran over, and only had thirty minutes to lay out a completely new timeline for the group. F, who we’ve been collaborating with on a data collection experiment, is rushing to take as much credit and as little risk as possible by early next week, and we’re going to try to have something more rigorous and rich to offer by 20 April as a sort of counter argument. So everybody’s existing priorities have been shaken down for loose change, and we’re running full-tilt at both the data we’re getting from F’s users and all the other data signals we’re hoping to include in our new tool. I spent the rest of the afternoon finishing up the public release of the crowdcast site and working on the F users survey data. First roadblock: CMU’s qualtrics license doesn’t include access to their API, so we have to download manually until that change request works it’s way through the bureaucracy. Okay; I switch over to the next stage of the pipeline. I take a quick meeting with R who replaces L with me as point of contact for data exchange on this project so L can focus on the statistical system. This is generally good because F is hella pushy and coping with that is not one of L’s strengths... but on the other hand if I don’t stop with this managing business I’m afraid they’ll just keep giving me power over things and I’m not sure I’m up for that, I really just want folks to make a fscking decision, do what they said they’d do, write things down, and communicate problems in a timely fashion, and welp I guess that’s management isn’t it. Next: F wants county-aggregated data ASAP, and we don’t have anybody on that, so I pick it up. I wound up working on that until... too late in the day for me, really, plus transition time to hand it off to L. Subsequent confusion about whether F wants counts or percents, and some anxiety about the percent-based HRR-aggregated data we already sent them before deciding to roll with it (F didn’t complain last time) and backtrack after delivery if needed. Finished around eight, then dinner, crossword, and sleep. I’ve already fielded a clarification question this morning from F, but I’m hoping they quickly get the picture that I am much more firm about boundaries than L was, and that sort of nonsense will trail off.

Yesterday I finished the elastic; another pair done. I went to therapy and talked about food issues, work, P’s work, and catching up with J. Shoved work issues under a giant rug made of recent progress, which she probably noticed but did not mention. Back at campus I struggled with lunch options and wound up just rage quitting and eating granola bars back at my office. Work was fine, currently working on backfilling some tests that were broken in the massive changeset from Wednesday. Meeting with M, his advisor, and F went alright but did put me on a later shuttle than usual. Home, sat for a bit, dinner was fine, crossword, Witcher, and a wee snack. I started reading The 7.5 Deaths of Evelyn Hardcastle by Stuart Turton. sleep. Dreamt of roof leaks and lab science and jelly beans and freezer burned blueberries.

Yesterday I had an early meeting on campus, so no morning hijinks. At work I power cycled my machine, and it seems fine now. Dubious dubious. I met with M to go over how the reference system works. I got lunch from the UC and it was okay. I fixed a pile of unit tests. I met with T to figure out a priority list for our current projects, which was good, because it turns out fully half of the semester plan from our last meeting is probably unnecessary, and the other half depends on the fate of software built by a grad student who will be defending some time this year. Whiplash, but not as much as it might have been if I’d spent the last month working on either project. I guess good job past me for procrasti-coding? Anyway, grumpy about that. Then we had lab meeting, where S presented his thesis work. He gives a good conference paper talk but has a lot to work on before he can cover a larger body of work as proficiently. Feedback was well-received though.

Home, dinner with a YouTube video on needlepoint lace, decompressed with P when he got back from his evening class, then direct to sleep.

Yesterday I made another batch of nut and seed bars. Library of the Unwritten has taken several unexpected yet delicious turns. At work I met with T, covered my current activities and the proposed next steps, and touched base about PTO approvals. I finished prototyping the new data structure retrieval bits and wrote a draft implementation that is now half debugged. At lab meeting O presented the ideas she’s working on for a possible thesis proposal about curriculum learning, which was fascinating. P drove us home, then dinner, crossword, Voyager, snack, and sleep. Long involved dreams about balloons that rose if you left them alone, but would sink to the ground for a game of chase; marching band practice (I’ve never been in a marching band but) conflicting with a surprise little league game, the worst logistics for a graduation of all time.

Yesterday I did my ballot research and voted. At work I started annotating the Inform7 code. I went to therapy and talked about my birthday and family health. I had a meeting with M and covered the inner workings of pytheo, and sketched out a plan for a LIA contribution relevant to his thesis project. I went to our lab meeting and heard G demo his poster for NeurIPS; it was excellent. Some visual tweaks but he’s got plenty of time before the conference. P had the car on campus so drove us home. Dinner, crossword, Elementary, snack, assembled ingredients for another batch of nut and seed bars, sleep. Vivid dreams of matches, parks, and recreational sod aeration.