Week #472
We diligently continued to work on surveys this month. One we’ve been running has taken a unique approach to their sampling. It is a parent survey meant for two parents. The most common options we’ve seen is to randomly select the first or second guardian. The downside to this method is that if one guardian just doesn’t answer, you loose 100% of your potential answers. Our improved method has been to split the survey into two parts and randomly assign the guardians each part. If you are a single guardian you get the full survey. Then, if someone chooses not to participate, we only loose a random 50% of the survey. Which is still better than nothing and it was more inclusive giving everyone atleast some chance of their opinion being heard. Now, we have a third method we’re testing. This randomly assigned one guardian the full-survey, but after two weeks, if they have not started, it reverts back to the other guardian and they have two weeks to answer. It will take a few weeks to know how well this method compares to the others, but we’re very interested in the results.
On the Material Conference front, we’re getting down to the more behind the scenes things, like confirming the venue, catering, speakers dinner, etc. We’ve also been teaming up with the local computer society, Ský, to offer their members a discount on tickets.
Week #473
This week, we finally got back on our newsletter horse and sent out a 02020Q1 update. It felt good to get back in the saddle and start to communicate via the newsletter again. As part of that, we’ve setup a new newsletter archive with some added metadata.
On Wednesday, we took our usual pair programming afternoon. What we expected to be an easy copy-n-paste of existing code turned out to be a sinkhole. We spent the whole afternoon unsuccessfully looking into getting the code to move data from one place to another. Some days it takes a lot longer to make that breakthrough. Luckily, we jumped back over on Thursday for 2 extra hours and managed to crack the nut!
Thursday night, we took part in the UX Nights local meet-up. We gave a short 20 minute talk entitled “The Web without a Screen”. We discussed previous projects getting the Internet on Paper, Voice Interfaces, interactive textiles, scraping and more. It was a lot of fun and good to start getting back into the local scene again.
This is also the week that the COVID-19 virus has started to take over in Iceland. More and more people are prepping, not coming out to events and worrying about the spread. A few cases have been reported in the country already and everyone is bracing for the worst.
Week #474
So many other events have been canceled or postponed now. Material is shaky, but we plan on proceeding. All the speakers are still planning on attending, but we’ve gotten a few drop-outs of attendees. Even if the audience is just ourselves, speakers and spouses, we’ll be recording and live-streaming the event.
That was written at the start of the week, by Thursday, we had to cancel. Not much more to say, but it was the safest, best option. Hopefully, we’ll be back soon with the same line-up of speakers.
We’re also back on VSK (VAT) returns every two months. So this week, we collected all our receipts, invoices, salary slips and other expenditures, collated them and took them to the accountant. These are the things that take time and are not directly billable to any client. So when you’re figuring out your hourly rate, you always need to remember to cover activities like this.
We also ran into some strange issues with Twilio. They have a Programmable Voice interface. You can initiate and receive calls and build a phone-tree menu using web hooks and HTTP Requests. We use this for robocall reminders in our survey tool. Then it stopped working, calls were not being put through. We could see in the logs that twilio made a connection and it was dropped after 1 or 2 seconds, yet our handsets never rang. Maybe it was our code, so we made a really simple example and tested an older one that we knew worked… still the same. We switched phone numbers and it worked again. Still some sleuthing todo, but we suspect the first number we rented has been black-listed, but no one has given us an answer of where yet. It is blocked across atleast 3 local providers, so it is not at an individual company spam filter. The company who provides Twilio with the virtual number says it is Twilio’s problem. Maybe after a few more support tickets, we’ll learn what and where the real problem lies.