Browsing through those large book stores, of the kind and size you can find only in the US, is one of my favorite pastimes.
I can loose myself in there for entire afternoons just letting the eye fly (wink) over the shelves and picking out the catchiest titles. My favorite section is of course the technology and programming one, and my first pass is often spent looking for the red spine and Newnes logo, just to make sure they do have copies of my books available. Too often they don’t, which I like to think is a sign of how successful my books are, they must literally fly (wink wink) off the shelves!
But for the last couple of years I have been attracted by books on Ruby, a very trendy programming language, although I never really ended up buying one or getting to dig deeper into the subject. I had this notion that Ruby was just for web site designers and there was little else for me there. Eventually, this Christmas I gave in and bought a copy of “Learning Ruby” written by Michael Fitzgerald and published by O’Reilly.
I choose it (among the many other Ruby’s books) because it promised to be very compact (about 200 pages) and based on a lot of practical examples, which is definitely my idea of the way books should be. Ok mine are thicker, but I promise, the next one will be shorter.
Fact is, I ended up liking this book a lot. I read it all in a single flight (from S. Francisco to Munich) and it gave me the sort of fever I sometimes get when something new gets my neurons all fired up at once!
Of course somebody has already started playing with the concept of porting Ruby to the embedded control world, with the RAD project (somehow it does not surprise me that this was done first on an Arduino platform, the A in RAD).
While I am not yet sure of the results and their application to the average embedded control project, I learned that “modern” interpreted languages are nothing like (bad) old MS-Basic. There are so many powerful concepts, such as reflections, duck typing, and multi-paradigm support to mention a few, that are worth exploring and considering carefully when trading performance (abundant in modern processors) with ease of development and code maintenance.
What is your opinion on interpreted vs. compiled in embedded control?