50th Anniversary of invention of disk drives 4588

At the same time, it's a shame that so many of the very clever algorithms optimized for serial storage seem to have been forgotten (or more likely never learned) by the last couple generations of programmers and thinkers. In fact the latest generation seems to be incapable of storing anything in a flat file, they now need a bazillion copies of Oracle everywhere, damnit!

In fact the concept of serial algorithms hasn't completely disappeared, but now it's called DSP and is practiced by signals engineers, not software guys in ivory towers :-).

And in an orthogonal dimension there are computing platforms where "run a program" or "access a file" are concepts that make little sense. Certainly the concept of a filesystem goes hand-in-hand with most uses of disk drives, but filesystems do exist outside of disk drives and disk drives are used in non-filesystem senses (here I'm thinking specifically of buttociative storage although there are other funky ways of thinking about how you get to the data!)

While the hard drive has made enormous leaps in just about every dimension (low cost, physical size, low cost, storage size, low cost, and did I mention low cost?) the one place where it hasn't progressed so much is access time. In the 30 years I've been in the business, CPU instruction times have gone from sub-1 MHz to multi-GHz (factor of several thousand) while disk drive access (latency) times have gone from 75ms to 5ms (factor of just 25). Most of those Oracle databases I gripe about are in fact cached in RAM somewhere :-).

Tim.