Just a decade ago the amount of memory for a typical desktop was in the lower 3 digits of MB. 15 years before it was in 2 digits of MB. Yet we had email clients, browsers, word processors, spread sheets, image processors. I read about one financial analyst’s perspective about the inflation and taxation in America. It is such that the purchasing power of money in your hand after taxes is significantly lower than the inflation adjusted money’s purchasing power in 1950. For example; if someone sells her house for in 1950 and then the amount of purchasing power that money had after the tax was significantly higher than what it could do in 2012. The direct and indirect taxes have risen and the inflation has masked the effects of it.
While writing this, I am doing so in a machine which has 8GB RAM, 500GB hard disk and i7 processor. That is a 250 times increase in my computing power than 15 years ago. Though the frequency of the crashes I encounter has reduced to a great deal, there are still crashes and non responding apps which annoys at times. I have observed that the thought “memory has become cheaper than efficient programs” is prevalent for most of the applications barring a few which demand performance. Take a look at the hardware of the Apollo Guidance Computer. We were able to land on the moon with that piece of hardware, but right now even with such excellent hardware we many times end up with non responsive websites. With a workload of just above 1000 concurrent users some web sites can easily choke the best hardware they are running in.
Is there a way that the generous doses of memory and computing power could be used effectively or will be waiting for something like water scarcity or environment pollution to happen to learn to use the hardwares efficiently?
Erlang it seems is the way to go, Erlang VM with a hand full of Worker threads can handle thousands of concurrent connections.