After a flu-break and some voluntary holiday related away-from-keyboard time, I’m finally back. I know, I still owe you the next part of the review of the book: The Craft of Research. And it will come, I promise. Actually, I read – while pushing our baby buggy through town – at least four (I might have missed one) other books: The Effective Programmer by Jeff Atwood, The Developer’s Code by Ka Wei Cheung, Designed For Use by Lukas Mathis and The Art of Woo by G. Richard Shell and Mario Moussa. The last one is still reading in progress, though. It is a quite tough read. I have the electronic version, and it feels that the real book must be a tome. As soon as I have some opinion about each of those and some spare time I will write reviews as well.
I even did get some more books as Christmas presents. Most notably: Cocoa Programming for Mac OS X by Aaron Hillegass and Adam Preble. I skimmed it a little, read the introductory chapter and feel my first OS X app will come soon. I will report and share it here. In the evenings, I was further playing around with my recent day-job project (which inspired my previous Xtend related articles: here, here and here). Despite my illness I do have a looming deadline to present some functionality working under real conditions. So the next post will cover another technical different aspect of this work. Basically, the integration part: how did I extended the legacy “closed source” web app – as closed source a web app can be, when you have admin access to the web app server – with custom information from arbitrary other sources. There will be some CoffeeScript, jQuery and cross site XHR hacking involved. But as you will see, not much magic.
You even might remember that I was fiddling around with the Olinuxino A13 and that there was some distant plan to build a robot with it. I still plan to do this and did put some thoughts in it. The first version will:
- have two driven wheels and a supporting furniture wheel.
- NOT be autonomic in any sense a robot can be.
- host a web server (node.js based) and allow to control the robot through a web interface.
- have a 12V 7Ah lead battery.
- use the XBOX Kinect for vision (RGB and Depth).
- stream the visuals through the web interface.
Thats basically it. I guess that should be possible and not that complicated. Furthermore, it looks like a solid base for sensory and actorly extensions and leaves the possibility to implement autonomous behaviour later on. Possible routes would include: mapping the rooms of our house and finding and connecting to its recharge station when power goes low.
And finally there is still my biggest project for 2013: my master thesis. I wrote a little about my plans regarding my thesis here. In the last days I thought about the basic assumptions again and must confess that I still have no real research question I think is worth answering and I’m thinking about changing the topic completely. You might remember that I looked broadly on artificial neural networks. Recently I thought about focussing on randomness and chaos in neural networks. But that might be too theoretical, as my master program is more business oriented. I will check back with my professor soon. My second thought was to take my day-job project as the basis and look at integration schemes to enrich arbitrary web apps with external information. A totally different topic and professor. I’m working on it and will report as soon as I made some progress.
In case you wonder, I will definitly not make a new keyboard in 2013! But maybe the following year and maybe I will make some modifications to Mark 1. If this happens I will as well report here.