I realize that I just broke my little corner of the internet by casually obliterating 14 years of blog content. 25858 files - gone. 4.7 Gb of content - gone.
As some of my avid readers may have noticed, there hasn’t been much content created in this space for quite some time.
The TimeExpander blog has evolved across multiple generations and flavours of content management systems over the years. At this point, both the site structure and site content has become heavily tainted with artefacts left behind in the form of legacy CMS idiosynchracies. On top of this, the page loading times has also become excessive.
Time for a change
Having left the “front door” open for 14 years without having been robbed blind doesn’t mean that it is a good idea to continue doing so. That said - the site never got hacked - which is a nice thing. I guess I owe my fellow netizens a big thanks for not attempting to gain entry without an invitation :).
The site now uses the Hugo static site generator.
Migrating from Wordpress to Hugo is possible, but I decided against it. 900 pages of text intermingled with theme formatting short codes and broken links required just too much effort to clean up.
As we all know, the open internet is archived - on the open internet. This means that it is entirely possible to browse sites that no longer exist - including the Wordpress version of TimeExpander that contains the content published between 2006 and 2019. Thanks to the nice people over at the Internet Archive and their excellent Wayback Machine, you will still be able to view snapshots of the old site >>> HERE <<<.
For the benefit of those who miss the slow loading times and who also want to reminisce about past content, I have compressed some of the less boring stuff from the previous years into a single 20Mb page (“Retrospective 2006-2019”).
Even if I haven’t been blogging much lately, I have been busy working on projects. Because of the hassles associated with the old Wordpress CMS, most of the progress on these projects has ended up being documented by videos or pictures on the Hackheim Facebook feed instead.
Having cleaned out the attic, I am now ready to start blogging again. Yay :)
In a couple of weeks time, I will be participating in a local robot fighting tournament hosted by Omega verksted in Trondheim. My robot, “Mostly Harmless”, is constructed mainly using plastic and composite materials. Since we are designing the bots while adhering to rigid weight constraints, I have spent most of my mass budget on “selected pointy bits and momentum building devices”. Where most people design in armour, I have only scaffolding, holding the main weapon in place. This “glass cannon” design most likely won’t survive for long, but in case something just touches the business end of this thing, momentum will be transferred and sparks will fly. Stuff will disintegrate.
The only thing now missing from my design is a weapon kill switch. I have absolutely no expectations about winning this competition (except maybe via walkover - in case I am the only participant…), but the fight(s) should be interesting to watch.
My nixie clocks received enough attention at OSF2019 to warrant making a limited run of kits. The PCB have been redesigned slightly in order to make the clock easier to assemble. This PCB version will contain through hole components only.
BBC Micro:bit minisumo kits
Autonomous robot building doesn’t have to be complicated or expensive. In addition to the Nixie clock kits, I have sourced the necessary components for a small run of BBC Micro:bit Mini Sumo robot kits. These will consists of a motherboard (for hosting the micro:bit), motors, wheels, motorcontroller, line sensors and infrared time of flight sensors for opponent detection.
This particular rabbit hole is terrifyingly deep and complex - and therefore in dire need of a TimeExpander travelogue :)
Further down the road
I am planning on getting my hands on some Dynamixel MX-12W serial servos. Having these in the arsenal would enable me to construct all sorts of robotic creatures from hexapods to bipeds. The MX-12Ws are among the cheapest in the Dynamixel range, but they are still rather expensive playthings. Especially if you need enough of them to be able to articulate all the necessary joints for a biped.
Combining the capabilities of the MX-12Ws with a tensor processing unit or two is way too tempting not to happen and being documented thoroughly on this blog - even if it will cost me the equivalent of half a pinball machine.