Week #449-450

Monday, September 30th, 02019 at 11:11 UTC

We missed last Friday’s weeknote because of the climate protests. So this week is a double weeknote!

Week #449

This week share’s its name with asteroid 449 Hamburga. This particular asteroid has a strange history on Earth. It has been targeted for potential landing several times. It is named after the German city of Hamburg, but McDonalds’ was interested in sponsoring a NASA space mission to rendezvous.

This week also share’s its name with Minuscule 449, a 13th century New Testament manuscript.

This week was mostly spent on survey tools. Finishing up a few translations. Mostly rectifying answer options between multiple sources. For instance, we have a massive dropdown list of countries born. That was available with the ISO code and translated country name for several languages. Now, we get a new translation language added, but their country code list omits several and added a few more. So it was a back and forth of removing a few from each set until there was consensus. It just goes to show that no matter how much tech or smarts you put into a system, it usually gets tripped-up by something completely orthogonal to the task at hand.

Starting in October, we’re reducing our commitment to these survey tools. They are mostly built and self-sufficient. The last few years have mostly been around getting them as anonymous and GDPR ready as possible. Now that’s done, it is time to move on. We had a potential project which was perfect. The timing was for 3 months, starting October 1st for exactly the amount of hours we needed. It was a lot of interesting, exploratory data visualisations for a company you’ve probably heard of. Nothing is every guaranteed and there might have been several people and teams pitching for it, but right at the end of the week, we heard it was off. They decided to do more internally and instead of a 3 month engagement it would be only a few weeks and/or maybe a training day. So, not completely bad news, but pretty disappointing and not the way you want to end the week.

Week #450

This week share’s its name with asteroid 450 Brigitta. It was discovered on the 10th of October, 01899. That’s nearly 120 years to the week! This asteroid is a main belt asteroid which is part of the Eos family. These are special asteroids that are thought to have at one time been one unit, but broke apart in a massive collision.

450 is also the phone area code for the Canadian province of Quebec.

This week was a bunch of behind the scenes work on the Material Conference. We’re started to approach a few local companies about shorter presentations. The event is still 6 months away, so getting commitment is too soon for locals, but we’re at least feeling things out and getting some reactions. It should be pretty exciting when things get a bit closer!

We also started to publish synopses of our previous talks. It has been really good to go back through all of 02017’s sessions, re-listen and take notes. On the day of the event we’re so busy, nervous and excited that everything hits you and washes over like a wave. Putting some distance between you and the talks and re-listening has been a great exercise. A synopses is written for almost all the talks now and they will be released one a week for the next two months or so. Then the holidays, and we’ll pick-up again next year with the 02018 event’s sessions. All leading-up to the next event in March 02020.

This week also continued a bunch of changes to the survey tool. Since we’re going to be reducing our time, there has been a last push to get as much done before October. We’ve imported old data, purged out other data, converted some encrypted semi-sensitive stuff and down graded it to regular values.

As part of many of our survey we offer a text-to-speech service. Typically, the work flow has been a bit convoluted but given that changes are pretty infrequent it has been OK. Each survey uses a yam syntax file, from that we generate two text files to be used in our text-to-speech desktop software. The first is all the phrases to be encoded into mp3s and the second is a map from the string to an md5 hash to rename the mp3s to make them unique by language and easy to reference in a URL. Since the software was only on one Windows desktop machine and the software is from 2013, it was time to move on. Luckily, Amazon Web Services offer a service called “Polly” which converts via a Web API text-to-speech. We spent a few days retrofitting everything and now we have an on-demand TTS server which speeds-up the whole process and removes the reliance on that one Windows machine and ageing software.

Then, after hours, Friday night we managed to finally, after many, many days, managed to have a call with some clients and friends in the US about a few projects. These are things that everyone wants to work on, but the timing obviously isn’t 100% perfect. But it is sort of now or never. So the really good news is that starting in October, we’re going to fill a bunch of our time on 3 projects. The Flynn project is getting a green-light in a much more formal way, but with a shifted focus. We’re also revisiting an old project again but this time for a new audience with a few more tweaks. Then the 3rd project is so early in development, this isn’t a code project, but more of a flow-diagram and slide-deck type of project. We’re really looking forward to these. Finally, we have the time officially carved out to work on these. The downside is that since the timing isn’t perfect, we’re going to run with this only for a few months, then regroup and see the status. Hopefully, we can continue, otherwise they’ll get put back on the shelf until the time is right.