#041

It’s the end of the quarter, so the pace has slowed, we’ve mostly been focussing on making sure everything is done and documented. Here are the highlights:

• Advanced search is gone, as regular search is now advanced enough. Doing search in only one way will let us simplify a lot of the finder-frontend code, so we can implement new things more easily.

• There’s a new facet to browse by supergroup in the search (click “+ Show more search options”). A supergroup is a collection of document types. We didn’t want to call the facet “Supergroup”, because that then requires explaining what document types are, so instead after a lot of deliberation we called it “Show only”.

• I started writing up the decisions made to update to Elasticsearch 6, but because we’ve not started serving search queries from ES6 yet that’ll hang around incomplete for a while.

Next week is firebreak, the gap between the quarters, and all the developers on GOV.UK are going to split into small teams and work on dockerising everything for local development, as the VM we currently use has issues.

I was thinking about integer linear programming (ILP), after being asked about making another rota generator at work. I’ve written before about modelling rotas with ILP. I was thinking about how it would be nice if the rota-generating code looked more like the maths. So I put together a little programming language for ILP (still incomplete!), which gets most of the way there.

I’ve been reading The Historian by Elizabeth Kostova, which is a horror-themed mystery about the legend of Dracula, which perhaps isn’t merely legendary after all. The story involves a few layers: the chronologically earliest character involved is an academic called Rossi; Rossi left letters and his research to his student Paul; and Paul is the father of the narrator. Each character got pulled into the Dracula legend in their own way (partly by their predecessor, other than for Rossi), and conducts their own investigation. The story is from the point of view of the narrator, sharing what she knew about the earlier characters and her own investigation. It’s pretty good so far, as of the end of part 1 (a quarter or so of the way in).

I’m investigating switching from Chrome to Firefox, and have discovered that they render fonts differently. Fonts are noticeably larger in Firefox. Bold is bolder. Glyph outlines are more jagged. Kerning just looks off somehow. These are all the case in both Windows and Linux. Not a great first impression, I have to say.