First create a bin directory in your home directory if you don't have one already:

mkdir $HOME/bin

Then edit your .bashrc file and add:

export GOROOT=$HOME/go
export GOOS=linux
export GOARCH=386
export PATH=$PATH:$HOME/bin

Then reread your bashrc file:

source .bashrc

You'll need some packages installed to build it:

sudo apt-get install bison gcc libc6-dev mercurial

Now check out the code and build it:

hg clone -r release https://go.googlecode.com/hg/ $GOROOT

cd $GOROOT/src
./all.bash

That's it.

The Sharp GP2Y0A02YK0F - Infrared Proximity Sensor Long Range -

Hal 1.0

While testing his motions, I was noticing some pretty erratic behavior, so I decided to collect some data on the sensor values on a fixed object, while Hal was not moving.

This is a graph of about a minute of Hal moving the sensor from 45 to 135 degress in front of him and then taking a reading when the motion of the sensor stops. It's at a wall, at a slight angle so the graph lines aren't on top of each other at about 45 centimeters away.

My results for this sensor are not within the tolerances of being able to program him to to anything really interesting in an environment that isn't really simple.

The good thing is that only once in this data would a "is this way further than that way" check be incorrect.  Then again, he isn't even moving positions here.  The data is pretty erratic over a 1 minute period.

In reality, this is maybe not the correct sensor for this application.  Maybe an ultrasonic would be better.  Hal may have to have eye surgery.

Hal IS supposed to be a very inexpensive robot based solely on Arduino.  I don't expect him to be able to do very complicated things, but I was hoping for a little better readings on a fixed object than what I'm getting.

As soon as I walked out the door to pick up our Chinese takeout for the evening, 100 ideas came swarming into my head. In my Jeep was a pen, but I cleaned it over the weekend and there was not a scrap paper in sight.

The moral of the story is twofold:

1. If you have writers block, a simple change of scenery can unblock you
2. Never leave home without some 3x5 notecards in your car

http://www.google.com/reader/wii

You add some people to help. They are a virtual team because the best people out there don’t have to move to be part of the team and you want the best.

Then you loose your original vision in feature creep.

18 months later, you have a bloated product that doesn’t really do anything and is 12 months from ever shipping.

Sound familiar?

I thought so.

I’ve done a number of startups. Some succeeded, some went down in a pile of flaming crap. All were worth it. Here’s what I learned. Okay here’s a few things of the 10 bazillion lessons I learned.

• To succeed you must lead these people you brought in. It’s your vision, keep them on your vision. They wont know how to do it. You can’t say “work on what you want”.

• No whining. If you whine, the people who work for you will whine to each other. And yes, they’ll be whining about you.

• People are not going to get it. You will have to cram it down their throats. Get a few poeple to LOVE it and they’ll tell all of their friends, who will tell all of… Once a few people get it, it may just explode into a viral phenominion that will have you calling me to figure your how the heck you’re going to scale this sucker.

• Keep it simple. Period. It doesn’t need that new feature for people to use it. Your sales guy is making an excuse because he doesn’t know how to sell.

• Stop having meetings. Or do what google does and limit meetings to XX number of minutes. Say it and get back to work.

• Lead from the front, not from the back with a whip.

Check it out at LondonFreelance. It’s in pounds, but you can convert.