{"id":490,"date":"2008-02-20T13:40:00","date_gmt":"2008-02-20T17:40:00","guid":{"rendered":"http:\/\/www.contrapositivediary.com\/?p=490"},"modified":"2009-01-14T23:23:55","modified_gmt":"2009-01-15T03:23:55","slug":"review-tom-igoes-imaking-things-talki","status":"publish","type":"post","link":"http:\/\/www.contrapositivediary.com\/?p=490","title":{"rendered":"Review: Tom Igoe’s Making Things Talk<\/i>"},"content":{"rendered":"

<\/a>Triage is a harsh mistress. I started out in computing with the CDP1802, a microprocessor designed for embedded systems work (it was used on the Viking Landers!) and for all the software I've used and the code I've written since then, I miss poking wires in breadboard holes and hitting the trigger of a wire-wrap gun. There's only so many hours in a life, and embedded work has not made the cut.<\/p>\n

That may change. Embedded systems tinkering is easier now. Much easier, and for a couple of reasons: 1) The processors themselves can be had on small boards with appropriate I\/O connectors; you don't have to fool with loose chips anymore. 2) Development software is better, mostly because now there is<\/i> development software. In 1976 I literally had to write binary code by hand. (F8 FF A2… Yes, yes, I know, barefoot and uphill both ways. But if you think I'm exaggerating, you simply weren't there.) And, the point of the current discussion, 3) there are books like Making Things Talk<\/a><\/i>, by Tom Igoe. Wow.<\/p>\n

The first time I saw the cover I was confused: It shows a stuffed monkey and the completely inane blurb, “Projects and ideas to create talking objects from anything.” I literally thought it was about fooling with speech synthesizers. But no: It's about networking embedded systems modules with technologies including Ethernet, USB, Bluetooth, and Zigbee. The microprocessor modules under discussion are the Atmel AVR-based Arduinos<\/a>, augmented by a host of sensor modules and connectivity modules that can be breadboarded on the same 0.1″ spacing blocks we used to use in 1976. In a sense, we now have modules the way other people have chips<\/a>. (And in saying so I am indeed just boasting.) We can raise our consciousness to the level of connecting functional blocks rather than individual inverters and logic gates. That is a very<\/i> big win.<\/p>\n

<\/a>The book is patient (as good tutorials must be) and begins with probably the finest introduction to low-level networking that I have ever seen. If you are a software developer you will understand it; if you have no experience whatsoever in networking or programming, you may have some trouble. A good prerequisite text would be Tom Igoe's 2004 Physical Computing<\/a><\/i> (written with Dan O'Sullivan) which focuses on the older Basic Stamp modules and their close relatives. Physical Computing<\/i> introduces both electronics and programming to a degree that Making Things Talk<\/i> cannot. That said, Making Things Talk<\/i> presents examples using a Java-derivative programming environment called Processing<\/a>, which is free and open source and much gentler conceptually than programming in raw Java or, merciful God help us, C. (Pascal would be gentler still, but as we all know, Pascal is a kiddie language that cannot accomplish anything useful. You must<\/i> believe this. A C programmer said it. QED.)<\/p>\n

Once the essential groundwork is done, the book teaches through projects, good projects that are mostly fun and in many cases even useful. The book explains how RFID tags work and how to read them, and how to read 2-D barcodes with a Web cam. The most fascinating projects are those that involve physical location sensing, using modules that perform infrared ranging, ultrasonic ranging, GPS, and “digital compass” modules, all of which made the robot guy within me itch like hell. We didn't have stuff like that thirty years ago.<\/p>\n

Although it doesn't get a huge amount of coverage in the book, the XBee module (which implements a Zigbee data radio system) fascinates me: It's basically jelly-bean logic implementing a short-range mesh network, and I intuit that hobby robots of the future may well consist of swarms of semi-independent functional blocks knit together coherently through the Zigbee network protocols, under the control of a multicore master processor. And damn, I would love to build something like that.<\/p>\n

Anyway. Here are a few additional observations on the book:<\/p>\n