Everyone who has anything to do with technology knows about Moore's Law. If you don't know it in detail or by name, you know it because you have a set of expectations about technology. You expect that whatever is available this year at a given price, next year you'll be able to get more of it and/or it will be cheaper and/or it will be smaller. This is Moore's Law in effect: computer-based stuff gets physically smaller, cheaper, faster; it can hold more and do more at the same price.
Moore's Law applies in spades to CPU's, memory, displays and even networking. All these things get amazingly better seemingly just by the passage of time. It even applies to disk storage: For example, the 1MB, 12 inch removable disk drives of my early programming days are now supplanted with tiny 300GB drives hidden somewhere inside my laptop.
While everyone more-or-less knows about Moore's Law, not so many people know about "Less's Law." Maybe it's because Gordon Moore was famous in his own right as a co-founder of Intel, while Seymour Less is famous only for confounding people who think that Moore's Law results in nothing but more and more good things happening. Seymour was fond of saying things like "the more Moore's Law expands disk capacity, the less Less's Law says you can access that capacity." In a time of belt-tightening, managers everywhere are saying "do more with less." Seymour Less is the guy who originally pointed out that, when it comes to disk, you have to find a way to "do less with more."
I've talked about the impact of Less's Law before, called it something boring like the "performance gap" in storage, and pointed out how it impacts the move to server virtualization. But I've been realizing recently that the implications of Less's Law go way beyond computer storage. The combined impact of Moore's Law (making most computer things better, faster, cheaper) and Less's Law (making storage less accessible) has a profound impact on software architectures. I still find ten-year-old software architectures being touted as "advanced," when they're anything but. On the other side, I see programming groups who are under pressure to deliver good stuff quickly adapting to the new world, and feeling their way to styles of building software that more or less reflect the combined impacts of Moore's and Less's Laws.
This is a big subject. I hope to explore it in future posts.
Comments