Tech Team Week Notes #44
For the ODI tech team, last week was "20% week". This is when we have time to ourselves to spend on innovation projects, learning, and so on. When we started, we tried to do a Google-style one-day-a-week rhythm for this time, but found rapidly that it was getting squeezed. Having one week in every five blocked out is a much more successful way of allocating the time, for us.
So, who got up to what?
Sam (who recently promoted himself to Head of Robots) spends his innovation time on projects that are fairly hard to tell apart from the rest of his work, spent as it is on automation of various types, but this week he put a lot of effort into bringing our test-driven infrastructure up to date with cucumber-chef 3. Apparently this is a lot faster and neater, so he's happy.
Stuart spent the week overhauling our tech dashboard. The old one had suffered some bitrot, so was due a few bugfixes. He also brought it into our main dashboards app, which uses Dashing, rather than using the hosted Leftronic solution which we started with. This is a step towards a larger piece of dashboard work we'll be doing over the next few weeks.
Ulrich spent the week working out tech demos for various projects and writing about the quantified self, followed on Friday by a visit to the Ministry of Measurement. He spent some of his 20% time working out ODI email statistics, but was "too successful"; too many automated emailsmade it a bit useless.
Tom spent the week partly in shock at the completion of our FP7 validation process, and partly working on various other project proposals. He's hoping to get some proper 20% time in this week instead.
Jeni spent the week responding to some bids, finalising proposals, and working on technical things to do with persistent resolvable identifiers (definitely Jeni's happy place, that last one).
Unfortunately, although 20% weeks are easier to block out, things do still happen that get in the way, and I spent most of the week chasing down a duplication bug in our invoicing system. It turned out to be bad HTML encoding in the hosted invoicing app we're using, but fortunately I could adapt our code to avoid the problem. Then, on Friday, I turned to innovation thoughts, and cracked out a quick demo showing CSV diffs in Gitlab, which I then wrote a big old blog post about.