PixelMEDIA Inc.

Last month, our Director of Engineering landed a few Google Wave invites and shared them with the team.  When I started playing with mine, I wasn’t even interested in talking with other people: I wanted to start playing with bots.  Google Wave is a terrific environment for bots and other automated tools, and with Google, they’re easy to code and deploy.

“Bot” is a term for any piece of software that can run tasks on the Internet.  Chatbots or chatterbots can carry on simple conversations or relay information to you; spyware bots can hack into your computer and plague your hard drive.  Used properly and legally, bots are convenient and helpful devices that can look up or process information for you.  Or they can just tell bad jokes.

For my first Google Wave bot, I decided to make a joke bot.  The concept was simple: you tell it a joke; it tells you a joke back.  The first version used a stock list of jokes that I had coded into it, but in the second iteration, I hooked it to a database and allowed it to learn and use every joke it heard.  This seemed like a good training exercise, and I managed to finish and debug it in about an afternoon.

Google makes the whole process very simple.  I started with these tutorials:

• Getting Started with the Google App Engine.  Your Google Wave bot will run on Google’s App Engine, which is free (up to a point).  The App Engine will host the app and its data, and it currently supports Java and Python.  I went with Java, and once I had installed Google’s SDK into Eclipse, I was able to compile and deploy my app with a single click.
• Google Wave Bots: Java Tutorial.  A step-by-step guide to coding your first bot. 

Now, as simple as this was, working with Google Wave presents other problems.  Google Wave is still not in a “final” release state, so you may run into bugs like this one, where everything I coded my bot to say was printed twice.  (Luckily, hunting around the forums led me quickly to this workaround.)  At this stage, you also can’t deploy and test them locally; once it’s live, you’ll have to balance every new feature you’d like to add against the risk of breaking the bot while people are using it.

But aside from these early-release quirks, I was impressed by how much trouble Google had taken to get people up and running with their bots – and I was even more impressed by the potential of these applications.  I used the App Engine to tell simple jokes, but you could build an application of far greater complexity into this interface – or connect to a solution that you host yourself, and that you coded in the language of your choice, and use the App Engine as a proxy into Wave.

The next challenge is simply to spread the word about the bot and get people to use it.  There are several public waves where you can test or discuss bots, although “spambots” – basically, any bot that speaks when it isn’t spoken to – are discouraged.  But I ended up using older social media tools like Reddit.com and Twitter.com to spread the word.  Wave is a powerful tool – but until more people get invites, the community will still feel small.

Tags: bots, Google Wave, Google Wave bots, Java