I have been finding microcontrollers and their ilk and derivations to be far more interesting than plodding around in a low-level language like C or even a high-level language (which I adore) like Perl. A very wise woman, Cheryl Mathews (if that doesn't work, try another), said to me one day that it seemed companies were fluctuating back and forth between scaling out and scaling up. It seemed that years ago you wanted giant racks of Onyx GigabillionRX-9's in your datacenter, and so soon, so very very soon, Sun (and HP and even Penguin Computing) were saying, well, we can outperform them for much less money, provided you spend a quarter million dollars on your myrinet/infiniband/craylink/yadayada backplane between machines. I'd never thought of it that way, but then I was really just a puppy when Cheryl and I worked together and I knew she was right. So I got to watching those trends.
It did seem there was a big bang/big crunch thing happening, and Cheryl seemed to be right. The testosterone in me wants that Sunfire 25k, and I liked the Netra X1 because it was a great price point. Cheryl took this in stride, in that miraculous way she has of listening to me, suggesting another (usually correct) example, without trampling on my toes that maybe this business of compute mountains vs compute rivers might all just be hooey.
Buy what we need, she said.
Cheryl was of course right. She seems to be usually right to me. :) But as virtualization came around, we realized that all that "mountain" iron would be sliced into little tributaries and gullies, wearing down the old salts that I aspired to be to the point They Saw The Light.
Is Cheryl a visionary? You betcha.
All this may come from nowhere it seems, but I am looking over the Android hackathon and I realize there is No Damn Way I am going, but if I know Cheryl, she at least knows about it or going herself.