Random Thoughts

Just noticed that my Oculus DK2 Postion tracker is not working

 and seems to be using a microsoft driver instead of the official one.

On the 5th December I took part in the Ludum Dare 72 Hour Jam. The theme this time was ‘Entire Game on one screen’. In the earlier voting rounds the theme of ‘Snowman’ had been popular and spent Friday night up to the announcement looking for ideas based around this. Then at 02:00 Saturday morning the theme was revealed and we had to come up with another idea. After sleeping on it it was decided to do Beat em up style game of children against Snowmen (I had found some pictures on Pinterest which we really wanted to use in the game) I was joining @Garethiw and @gredgie , The plan was that I would come up with the sound and other areas that were normally rushed near the end of their previous Jams. As Eddie Izzard put it  when talking about his Clarinet playing in a band, I would be doing ‘the bits that were left over’. So we had a theme and off we went. Garethiw wrote all of the game code in Unity and Gredgie created the graphics, while I came up with sound effects, some of the UI and the bodycount scoreboard scripts. Our work was committed to Github which seamed to work well apart from some issues with prefabs. I had great fun over the weekend especially when creating the voices for some of the creatures in the game. As with my previous entry I marvel at the ability of my friends and that they let me join them. Time lapse of the weekend

Over the weekend of 22nd August to 25th August I took part in the 30th Ludum Dare Game Jam. The aim of the jam is to create a game within 72 hrs based on a theme revealed on the Ludum Dare Site. The theme this time was Connected Worlds, we had been coming up with rough outlines of ideas for the final 20 themes and this was the one we didn't want (always the case !). After some discussion we decide to go for a Romeo and Juliet based game where you were trying to defuse the hostilities between the 2 families result in the main character not killing themselves.

Scala runs on the Java Virtual machine (JVM) and compiles to byte code the same as Java, this means that like jython it can also use Java libraries. One of the nice features of some modern languages is the REPL a console that allows the direct input of code to the interpreter along with instantaneous feedback.  PTC's Windchill PLM/PDM Application is written in Java and also features an API to allow it to be customised to meet customer requirements. Running the Scala console inside a windchill shell allows access to windchill libraries and hence allows code be run on windchill's Method Servers using remote method server calls. To give it a try download and install Scala from http://www.scala-lang.org/download/ then open a windchill shell and type scala to access the Scala REPL. Like Java, before you can make any calls using the Windchill API you need to import the classes, also to be able to run windchill methods we need a remote method server instance, so we can enter

import wc.method.remotemethodserver

 followed by

val remotemethodserver=RemoteMethodServer.getDefault()

 now that we have a method server  we can access it's methods, by entering

println(remotemethodserver.getInfo().startDate()

we can print out the start time of the method server as shown below.

Typing this in line by line would soon get repetitive and boring so we can create our code in a text file and then run it via the Scala console using scala filename.scala in this case scala WindchillExamples.scala

The below code shows some examples of using the windchill api to create a part and query the database, the syntax looks similar to how it would be written in Java with a few differences (specifying the class in the query spec and casting the returned part object from Persistable for instance). 

I am looking to try and run a workshop session at my local developers group, the group contains developers form a range of background and hence a number of languages, this makes workshops interesting. To be able to do any interesting projects with a Raspberry Pi I would prefer people are spend their thinking about problems and not battling which ever language is currently running on my RPi devices. A recent pole suggested that javascript was the most widely used language among the group so I decided to see could be achieved in Node.js. I found a few libraries/modules that allowed gpio access and decided to try onoff since it the example code included using an event to trap button pushes. I am not quite sure currently what my final project will be but starting off with a simple project I decided to try and control my Berry Clip 6 LED addon board. The below code selects a random LED to turn on and turns off the previously lit one. Full Information and a run script can be found in the README.ME

var Gpio = require('onoff').Gpio, ledA = new Gpio(04, 'out'), ledB = new Gpio(17, 'out'), ledC = new Gpio(22, 'out'), ledD = new Gpio(10, 'out'), ledE = new Gpio(09, 'out'), ledF = new Gpio(11, 'out'), button = new Gpio(07, 'in', 'both'); var leds=new Array(ledA,ledB,ledC,ledD,ledE,ledF); var lastLed=ledA; button.watch(function(err, value) { var activeLed=Math.floor(Math.random() * 6); lastLed.writeSync(0); lastLed=leds[activeLed]; leds[activeLed].writeSync(value); });

For a while now I have wanted to try using analog inputs with my Raspberry Pi. The Pi hardware only supports digital inputs, but it is also capable of communicating with expansion chips using an interface called Serial Peripheral Interface (SPI) . I decided to try my hand at getting a ThumbStick from a GamePad to work They are available for a few pounds from ebay and so would not require me to destroy an existing console controller. 

Some internet research pointed me in the direction of using a MCP3008 10 bit Digital to Analogue converter. The first guide I found for using one of these was on the Adafruit learning system and used something called Bit Banging, further research showed that later Pi do not need to do this and a guide on Raspberry Pi Spy gave me a starting point using a library called py-spidev. This lead to the below circuit diagram. The 2 direction (X & Y) signals are fed into the A/D converter. When the AD converter is queried by the python code a value between 0 and 1023 is returned representing the position of the stick ie 0 is extreme left, around 500 centre and 1023 extreme  right.

The next part was to make my stick work like a joystick. I decided to have a play with PyGame (PyGame is installed as default on the Raspbian image) and see what I could come up with. At IWDEV 2 Gareth Williams gave a Introduction to Monogame and created a simply shooter based on an alien and his cat. Trying to recreate this in python on the Pi would provide me with a suitable test for my joystick.

Once this was working I started to create some code to act as a bridge between the A/D and the normal Joystick functions. Luckily this proved to just be a case of setting range of the A/D converter values when initialising the uniput library and then passing in the value received by querying the chip. I used a third analog channel for the switch activated by pressing down the stick but it should really be a digital GPIO input and I was being lazy. 

The results as a Google + Auto Awesome montage

The python uinput library download and install details can be found at:

http://tjjr.fi/sw/python-uinput/#download-and-install

Matt Hawkins guide to the MCP 3008 including enabling the kernel modules and installing py-spidev can be found at:

http://www.raspberrypi-spy.co.uk/2013/10/analogue-sensors-on-the-raspberry-pi-using-an-mcp3008/

My final python listener script can be found on github at:

https://github.com/qubecad/Pi_stick/blob/master/analog_joycestick.py

To start,  run the .setup.sh script which loads the uniput module and then start the python script with the command:

After a recent visit to see my niece and nephew I wanted to see if I could come up with an app that they might use. My idea was to create a paint by numbers style app but to source the picture to be coloured from an image.

The result is Snap N Paint, the picture is captured via the camera and processed using the OpenCV library sobel function, this provides an outline which can then be coloured in.

To make sure the correct colours are available in the palette, the original image is used. The user selects the area they would like to colour and the correct colour is set active.  

I saw a recent post on Raspberry Pi Spy about the Pi edition of MineCraft having a python API and thought it might be fun to play with. After installing, it seems it also has a Java Api which closely follows the Python. The next question was what to write, always a sticking point. I decided upon flipping things on their head a little and have a joystick in MineCraft control my trusty PiBot (I've tried to control it with everything else why not this). Hitting one of the blocks via a right click would launch the Bot into action.

The Pebble 2 way Communication SDK for Andoid and iOS has just been released, so what better way to try it out than by adding Pebble controls to my Raspberry PiBot Controller.

Pressing up on the Pebble steers left, select is a toggle between forwards or stop and down steers right. The web service on the Pi returns a Json status message which is displayed on the Tablet Console. The tablet app also features voice commands via Android Google's android services.

This is also repeated onto the Pebble when a command is sent from either device.

On Sunday I took part in Walk the Wight 2013, a 26.5 mile sponsored walk in aid of Earl Mountbatten Hospice. The route takes in most of the length of the Island incorporating the Tennyson Trail and includes some wonderful views of the island (weather pending).

Walk the Wight Route plotted via RunKeeper

In keeping with my usual habit of  using gadgets in everything I do I decided to try and record my progress via one of the fitness apps available on my phone. I backed the Pebble smart watch on kickstarter.com and one of the first apps to support it is Runkeeper so that was a logical choice. Runkeeper tracks the route, speed and altitude and lets you view you result via a web portal (the Runkeeper results can be view at http://rnkpr.com/a2ywwyb). The Pebble has it's own application running on the phone managing the blue tooth connection and providing some notifications, this can be further extended using Pebble Notifier which relays notifications displayed on the phone to the pebble so any tweets or SMS messages I received during the walk were displayed on my wrist instead of having dig my phone out of my bag. This was especially nice in the last few miles of the walk when the weather took a turn for the worse and it was nice to get encouraging messages coming through from friends.

Pebble Watch showing data from the RunKeeper App running on my Nexus S

Keeping the GPS running on a phone is a sure way to flatten the battery, so keeping it running for 9+ hours of the walk was going to be an issue. When I was building my Pi Bot project I used a Pebble Smart stick as a power supply, so this was ideal to extend the power to my phone. I have recently switched to using a LG Nexus 4 but it is rather power hungry compared to my older Nexus S so I decided to use my Nexus S instead to further improve the power consumption. I raised around £120 for the hospice and for most of the walk the weather was sunny and enjoyable even if the last bit did end up rather wet. This was the third time I have taken part in Walk the Wight and every time it is wet I say I am not doing it again but I suspect once my aching muscles have recovered I will think about doing it again sometime.

Finished and you even get a medal!