Re: relational reasoning -- why two tables and not one?

Regarding vocabulary, I read somewhere that while there are more than half a million words in English, the language with the most words, it is rare for one person to have a vocabulary of more than thirty thousand and most people operate with fewer than ten thousand. I have no idea by what means these were counted, but I'd guess that if we throw out the prepositions, adjectives, adverbs, conjugations and so forth we might be left with a couple of thousand nouns of which only a few hundred are used every day. Being cynical by upbringing, I'm sure that number is declining, certainly it is among the rap generation. (For database purposes, I assume we can throw out all but a handful of verbs since computers are incapable of enacting much more than the boolean logic operators.)

I remember when the large canned db apps came out, about twenty years ago. They were described as 'enterprise-scale', being successors to the specific narrow apps such as payroll, ap, ar, gl that had been around since the early 1970's. Never had anything to do with them but I gathered that they fell out of favour partly because they usually needed massive customization that could be more expensive than "rolling your own".

I doubt if I'm going anywhere useful with this but I've often wondered just how many data relationships an efficient economy really needs. I can vaguely remember the early 'data dictionaries' from the 1970's (usually they referred to non-relational data idealogies). Maybe somebody can help me out as to what they contained, I seem to recall there were many verb-phrases in them, which seems silly in retrospect.

Even if the average business needed a mere one hundred nouns to encompass its operations, I guess that could still entail about two-times-something-raised-to-the-power-of-a-hundred possible relations, a very large number, intractable as they say, whereas I believe the very largest databases are likely to number their relations in the hundreds.   But what if there were only one hundred base relations and no user ever saw anything but views?

Maybe my question, if there is one, is this no doubt naive one: Why are there so many different db's in the world? Is it for the same reason that no two automotive marques use the same cranks for opening their windows? I started to think about this when I spilled wine over my little Linux laptop and apparently it has a permanent short-circuit. For three-hundred bucks plus tax I found another laptop which has this Vista OS on it. For my purposes the only difference is a bunch of gew-gaws that periodically get in my way but which I can tolerate. I gather that some people think it has better integration and some people point out that certain interior functions have become slower and 'mpg' has gotten worse but it mainly seems to me to be yet another window crank design. Received on Fri Oct 16 2009 - 01:13:54 CEST