Week #680 & #681

Friday, March 1st, 02024 at 13:31 UTC

Experience it 3 times, once when it happens, again when you write it down, then a third time when you read it again. Sometimes we need to remind ourselves that these week notes are also a reminder of what we did when we look back in a few years time.

Week #680

On Monday, we had an Unoffice hours time slot booked with the cryptic discussion topic of: Let’s talk about what words should and shouldn’t be used. This got our mind racing since we’ve been working on the dictionary work on SpellStruck. Was this a disgruntled player? Had we said something someone took offense too? It is interesting, like a horoscope or psychic medium, the messages are vague, but not so vague that you couldn’t apply them to something you’re thinking about.

Well, they never showed-up. We sent them an email with no reply. Maybe something happened, maybe they’ll reply and maybe we’ll reschedule. Time will tell – probably spam.

We also went down to the University of Reykjavik, where the staff in the bio-mechanical engineering department are helping us to test our prototype bike design.

We have some data from the tests, but need to do another round and investigate some other aspects.

On Tuesday, we published our 02023 Annual Report. This is a high-level overview of where our time and money went. It is meant for others who are freelancing (or considering it) or running a small business to compare and contrast to see what goes into running a business. It’s easy to have an idea and follow your dream, but few realize how much time, effort and money is spent in the infrastructure portion of a business – taxes, salary, accounting, legal, office supplies and more.

We continue with two projects in Australia. The first is glueing two APIs together. That’s going well, we have a domain and hosting, now it is the workflow and what data they need out. The second is cleaning-up and adding more features to our web-interface over Xero accounting software. The company can login, see active jobs and archive them. When they do, it creates an email with all the P&L info they need.

We have two on-going parent surveys which are well underway and in great shape. In March, we’ll start a staff survey. Easter is the last week of March, so we need to start on time and power through the first three weeks. This week we continue to collect lists and prep. We also will start a third parent survey at end of Feb which will also run in March. There is a little bit of work we need todo to prep for some longitudinal changes to better optimize the sampling methodology.

Week #681

This week we wrapped-up two big parent surveys and got ready to start a third. We’re also planning on starting a staff survey in March. It’s in good shape, we just need a few more lists, but given Easter is at the end of March, we’re in a slightly compressed timeframe. We normally give a parent 10 days to answer before swapping, so 21-ish for a staff member is plenty – even if we loose a few days to a slow start.

On Monday, we had our fortnightly PETALS meeting. We are at the point where we go through the whole flow and find less and less bugs. Some small clean-up this week, then the plan is to get an external team using the system to get more feedback. It’s an exciting baby-step to making this more public and a proper product.

We sent off another print-on-demand project idea. We got the proofs back already and have handed a few out for people to test. This is a riff on a popular genre of word-puzzle game where you get a quote or phrase but all the letters have been replaced by numbers. Then the key gives you a few of the pairs and much like a sudoku, you need to figure out the rest through logic and trial and error. After the feedback we can tweak the difficulty a bit, find more phrases and iterate to something more public.

Our office space is slowly coming together. We moved to a new, larger room, but in the shuffle, it became a storage room and we’ve been Tetris’ing around it for a few months now. This week, we got the shelves assembled again and books have been organized. That got us another few square meters and we have managed to bounce back and forth to something that’s almost usable!


The BIG project for last two weeks has been a project which requires video chat. Rather than re-invent the wheel, we implemented a WebRTC client on the web and in iOS. WebRTC is a bit like magic because it is built into modern browsers. You do a lot of voodoo to make two (or more) clients talk to each other, then once you’ve negotiated the streams, you basically point it at a <video> element and it just works!

We found some python libraries and javascript examples along with the same for iOS. But the only part of the WebRTC spec that is designed is the peer-to-peer video connection. How you signal to each other is up to you and of course everyone does it differently!

The Robustness Principle states: “be conservative in what you send, be liberal in what you accept”.

The code now has lots of if type==”ICEcandidate” or type==”IceCandidate”… seems like every library does it slightly differently. This was the prototype to prove to the customer it is possible. Now, if they are interested, we can work with them to build out the bigger concept.


Back in 02019, we wrote our #420 weeknote. We were still deep in survey work at that time too! But we also mentioned the new “TV Game” Avo! Sadly, the game and studio PlayDeo no longer exist… probably because Apple hired away two of the three co-founders in 02022. It was a fun game. You can find plenty of walk-throughs up on YouTube.


This week, a friend introduced us to the YouTube channel 3Dbotmaker. They have die-cast car race tournaments. They commentate as if these were real sporting events with plenty of background, driver history and more.

Another VERY similar concept is Jelle’s Marble Run channel. They have obstacle courses for marbles with plenty of teams and back stories. It’s fun to watch and you quickly get caught-up rooting for your favorite team.

Speaking of Marble runs, there are channels (Mr Sticksman) using virtual marble run software to build massive obstacle courses. One of the interesting ones has a patreon where you can help support the channel. In return, one of the virtual marbles gets your name on it and you can watch and root for yourself. We think that’s a great idea, it is small effort for the creator, but a big, personal impact for the contributor. It is also virtual, so the cost to label the marble compared to what you can charge has a pretty high mark-up compared to making something physical and shipping it through the post.