A local company is letting me test run a 4-hour TLA+ workshop next week, so I’m going to be spending this week writing it. Otherwise, more setting-up-business stuff, prolog class, and biking.

Finished the user end (Swift) and the backend (Common Lisp) of my first ever iOS app. Now looking into enabling remote push notifications via APNs. Unfortunately it’s HTTP/2 & ALPN only, and there’s no Common Lisp library to handle that directly. There’s an H2-14 reference implementation which unfortunately does not support ALPN and have suffered from some serious code rot. So my choices are either fix that up, or get it via some existing library (libcurl?).

I am updating Merit’s ambassador reward and invite distribution algorithm. The community voted on changes they wanted to the algorithm, and I am implementing them in the Merit Core software. It’s quite exciting to work on a decentralized software project because we can’t dictate that users will upgrade the software unless they have a good reason. Since the changes we are making are based on what the community wanted, then the likelihood of adoption is pretty high. Also unlike many different projects, the new features are turned on essential via a timed feature switch. So we must deploy the software with the feature disabled, and it automatically is enabled at a future date. Timed feature switches are necessary because most people on the network must turn on the feature at the same time for it to function, or else you get a hard fork in the blockchain.

At work we just released a dataset that contains 18 years of hourly air quality levels in Madrid in Kaggle, so now I guess it’s time to play with it! I spent 2 days cleaning the data from Madrid Open Data, so I’m also writing a post complaining about how much you can obfuscate data and still call it “open data”. Last but not least, I still have a master thesis to work on this summer!

• For $CLIENT1 I’m aiming to actually start the work to migrate from ZF1 models to Bamboo v5

• For $CLIENT2 some more minor tweaks to the data capture system

• For $COMPANY I discovered last week that the current time tracking and invoicing software we use is dropping support for self hosted instances as of early 2019, so I need to either find an alternative (realistically needs to be open source) or build/adapt one in-house.

• At $HOME the builders have finally started on the renovation. It’s amazing how much you can communicate with a few words of each other’s language, and a fuck load of pointing.

Work - all over the shop, mostly some work on replacing some existing crypto key management with Vault’s transit backend. Also helping our QAs use the system architecture to make decisions about test design and coverage.

Non-work - I wrote a side-project app to keep notes on research papers in GNUstep, and used Renaissance for the GUI (it’s an XML format on disk, and an easy layout system, very similar to the Stack Views stuff in very new iOS). It’s missing support for some view classes that I’ve been laying out in code, so I’m going back and adding that to Renaissance.

Additionally I recently started documenting my views on OOP, having spent years doing talks all based on the premise that the easy bits are useful, the complex bits are generally wasteful, the whole thing has become internally inconsistent but eventually you have to give up trying to write procedural programs with “classes” as the units of modularity and start composing objects.

Non-computer - a choir I’m performing in for the Warwick Folk Festival has its first practice this week.

At work, more tickets, and trying to find process and performance improvements.

At home, getting ready to move halfway across the country, from Florida to Missouri

Two days of team all-hands today and tomorrow - great to see everyone but we all seem to think two days is overkill (except our director).

Head to toorcamp on Wednesday though via an early ferry! I’m excited for the weekend hacking in the woods with friends.

At work I’ve been triaging new cases in the lead up to our next release and working on some new statistic collection APIs for the next next release when I get a chance.

Outside of work, I’ve been super busy.

First, I’m taking this “Introduction to SWI-Prolog” course. The first week was a basic intro to SWI Prolog, setting up our environments, and a brief overview of the built-in editor and debugger UI. The upcoming week looks to be more interesting, as we’ll start start digging into the language itself and writing some code. The instructor is definitely excited about the language, so I think it’ll be a fun course.

I’m also “training” for some bike tours I have coming up in July. The 1st - 7th some friends and I are going on a hut trip in the area of Colorado between Eagle and Leadville. Later in the month a different group of friends is participating in the Steamboat Ramble between Fort Collins and Steamboat, and I’m taking the following week off of work and biking back home to Boulder.

And, more excitingly, I’m planning to take a year off of work starting in October. So far the idea is that I’ll spend the first six months in a cheap apartment in a ski town (cheapest I can find, anyway) and relax and ski all winter. Then, in the spring, I’ll take off on my bike and tour around the western United States. I’ll use the whole year to catch up on my enormous reading list and contribute to some open source projects.

Last week I finished my article for Neon Dystopia about the roots of bOING bOING (the zine), https://www.neondystopia.com/cyberpunk-books-fiction/the-brain-mutator-for-higher-primates-a-boing-boing-retrospective/

This week I’m continuing my documentation for emulating a z/OS mainframe and then hopefully documenting my PBX configuration.

Another week of being all over the place. Working, Sailing, heading to NWRUG (Ruby user group), Gardening and no doubt working on the car.

Also need to put two cars up for sale, turns out owning four cars is hard work and I only really need two in the family. Oops.

At $work, I’m going to plug our Graphite setup into the in-house instrumenting pipeline so we can detect new bugs more quickly than by people pinging us asking “Is the thing broken now?”, and then get on a rollout train. In the medium term, I also want to hook us into the in-house deployment setup so we can get faster rollouts (but mostly faster reverts) than we currently do via puppet, but that’s going to take some sysadmining and sysadmin-convincing.

a toy type system, hopefully getting things solved to get back on track with college and life after a hard couple months