Re: Hashing for DISTINCT or GROUP BY in SQL

On 15/10/2010 10:45 AM, Anne & Lynn Wheeler wrote:
> I was undergraduate at univ. where the library got ONR grant for doing
> online catalog ... part of the money went to getting a 2321 (to host the
> online catalog). The project was also selected to be one of the original
> betatest sites for CICS product ... and I got tasked to support/debug
> CICS.
AIR, it came out around 1965. The big news was its capacity, 400MB! (no, I don't mean factorial, that's all it had!) Some people called it the 'spaghetti picker'. The typical computer room raised floor might vibrate when it was in operation. At our hole-in-the-wall industrial unit, we had it on the concrete floor and it would perceptibly bounce across that floor a little bit when it was going full blast. (The two and three inch diameter power cables I had snaked all over that floor kept it from shuffling out the door.) But at the time, the 2314 disks came in a bank that had 8 (plus one spare) drawers, about twenty feet long, three feet wide and five feet high. Each of those had a max capacity of 29MB so the 2321 had about 50% more space, albeit with an order of magnitude slower seek and access time.

In the 1970's IBM came out with 3330 and 3380 disks that had somewhere between 100 and 300MB capacity. Nothing compared to today but it might explain why the original IBM PC hard-disk of only 10MB was initially thought to be quite acceptable.

By the 1980's it was common for programmers at many sites to never see the computer room. Once, a younger programmer asked me how big a disk drive was. I told him it was about the size of a large cake tin but that it was housed in a much larger box so the the customer engineer could get his head inside. Indeed there was an IBM CE who was decapitated while servicing a similar device.

The early 'databases' used by the first wide-reaching online network systems such as PARS involved ruthless reduction of 'attributes' as it were. Yet it was still possible for the real-time flight and cargo 'databases' of large airlines to be stored in some hundreds of MB's, in other words they could be recorded in the main memory of pretty much any of today's consumer pc's. Had such memory been available thirty or forty years ago, I'd venture that the programming landscape would look different today. There remains much bowing and scraping towards legacy obstacles. From papers they wrote, it looks like even the System R people's thinking was dominated more by past physical history than the likely future. Me too, it wasn't until the 1980's when I could actually put a computer under my arm that I started to realize how much more important the logical side of programming is. I think younger people are hide-bound in a different way, there are now so many different programming languages and therefore idioms which encourage them to think that all that can be invented has already been invented. Received on Fri Oct 15 2010 - 20:22:48 CEST