Luke Maurits

I relatively recently learned (I don’t know when this first actually became true) that you can now have arbitrary collections of Wikipedia articles printed into actual physical books, which can give a title to and design a cover for, using images from the constituent articles.  Furthermore, the prices are relatively reasonable (and you can download pdf versions of the books for free).

I’ve been intrigued by the possibilities of print-on-demand book publishing for a long time, and also a fairly obsessive Wikipedia autodidact for a long time as well, so when I discovered this I went into a little bit of a frenzy, where over a period of 2 weeks I composed books on just about everything I’d like to know more about – linguistics, genetics, evolution (and in particular human evolution), neuroscience, Earth sciences, and more.  I absolutely couldn’t afford to get all these books printed (and shipped – shipping costs make the prices of these books not quite as attractive), but to test the service out ordered my linguistics and genetics books.  Including postage they cost me about AU$30 each.  Not super cheap, but given how expensive “real textbooks” are and considering that this system gives you total control over what material the books cover, certainly not bad either.

It took quite a while for them to arive (probably something like 3 weeks), but I was fairly happy with them.  The printing and binding quality seems quite good – the spine of my genetics book was a little crushed at the top, but that’s probably the result of rough transit experiences rather than a manufacturing defect, because the bottom of the genetics book spine and both ends of the linguistics book spine look great.  The system they use to make the books is based on LaTeX, so the pages look quite pretty, although there are some flaws – clearly they don’t convert ” characters into ` and ‘ characters depending on their position, so all the quotation marks look a little bit off.  This is a fairly novice LaTeX mistake and I have no doubt they’ll change it pretty quickly.  It would also be nice if you could choose to have the references at the end of each article excluded.  It’s good to have them there on the Wikipedia website, but in book form having a few pages of hyperlinks at the end of each article is a waste of paper.

The really big issue around these books is just how coherrent a text you can get from concatenating a pile of Wikipedia articles.  My linguistics book was a bit of a grab-bag by design, but I put a lot of effort into planning the genetics book.  You can arrange the articles into chapters, and before getting the book printed I ran my chapter/article structure past a geneticist friend to make sure I was covering the major subjects and arranging things in a sensible order where articles wouldn’t depend too much on information in articles that came after them.  The end result obviously wasn’t as “smooth” as a real genetics textbook (but it was an awful lot cheaper), and there was quite a bit of repetition (including some verbatim repetition where text had clearly been copied and pasted from one article to another), but at the same time at no point did it leave me really confused and scratching my head.  The amount of biology background knowledge I bought to the task of reading the book was very minimal, and I definitely came away from it with what I think (and hope!) is a fairly solid understanding of the basics of the subject (I was able to follow this recent public lecture without any difficulty at all).

All up I’m pretty happy with how the books turned out, and will definitely gradually order the other ones I made.  It might seem silly to some people to spend money on printed copies of material which is freely available online, but I think having a big bunch of related articles in a book you can read on the bus lets you really learn a lot more from Wikipedia than idly reading individual articles online as particular things take your fancy.  It also solves one of the main problems with reading Wikipedia online.

It’s been a while since I first posted about my plans to build an ROV, and I’ve been quite silent about it since then, but progress has been slowly happening.  I’m hoping to get a chance to do a lot of construction work done in the next week or two and I’ll make sure to post pictures of as many stages in the building process as I can.  I have the majority of the frame cemented together now, and it turned out much better than I’d hoped.  PVC pipe and the standard fittings for it obviously aren’t intended to be used to build complicated, precision structures, and I’d worried that I’d be left with something a little bit wonky, but it turned out great.  Both the top and bottom parts of the frame are almost perfectly square, are the same size as each other, and really quite tough.  I’ve bought a single 450 GPH bilge pump to “test the waters” with (buh duhm kssh!) and I hope to be able to post a YouTube video of a simple bathtub test soon.

During the time I’ve actually been working on the frame, I’ve been thinking about the (to me) more interesting and exciting stuff, like control.  My original plan had been to keep things extremely simple, and just control the thruster pumps using a block of relays which I sent signals to directly from a control unit on the surface over some ethernet cable.  However, I’ve been doing a lot of reading lately on the Arduino boards and how to do all kinds of things with them.  They are much, much easier to understand and work with than I ever expected and I pretty much can’t wait to buy one and start playing with it.  So I think I’ll put an Arduino onboard my ROV from the start – to begin with I’ll just use it to control the pump relays, but I think I’ll pretty quickly want to start attaching devices to it for navigation and recording data on things like temperature and water pressure.

One really nice thing about using an Arduino for control is that I can put an ethernet shield on it and then control the ROV directly from a laptop, rather than having to wire up a box of joysticks and/or buttons to do control.  I had intended to have a laptop present for ROV sessions anyway (to display camera footage), so this actually reduces the amount of work involved to some extent.  Furthermore, if I can find a cheap IP camera to use instead of a USB webcam, and I can find a very small ethernet hub or switch (this page suggests it may actually be easy to build such a thing using nothing more than some diodes!) I can get away with the one ethernet cable for control and video.  My original plan had been to have one cable for talking to the relays and a separate one for video, which would carry the signal from a USB webcam using an ethernet-USB extender.  Cutting down the number of cables in the tether will reduce cost and increase speed and maneuvarability, which are both good things.