Re: Entity and Identity
Date: Wed, 22 Jul 2009 21:46:59 GMT
> with a module system. Procedural can exist quite happily with
> declarative - take a look at the procedural interpretation of horn
> clauses, and continuation passing style programming.
The popular OO programming techniques/languages commonly depend on assignment, no matter how much they pretend otherwise. For the ones that don't, there might be cases where I'd grant your point. A procedural interpretation on top of a declarative system is a different dimension from one that is logically equivalennt, a bizarre hybrid, like mules who can't reproduce and comparable to nothing and everything at the same time. The clause perspectives you're talking about could be interchangeable in some settings and one might not need to depend on both at the same time in order to understand the intention or get results. Whereas the OO-relational silly conciliators basically pretend that a relational scheme is nothing more than a file system. This is of the same uselessness as EAV because it is fatuous to build something you already have. It amounts to accepting unneeded complexity for the sake of 'jobs for the boys'. Good cooks know to never combine more than three basic flavours, in programming it is ideal to keep that down to one flavour, same reason as a man with two watches never knows what time it is (and this ignores the general incompetence of most programmers, which is another story entirely).. These comparisons may seem like obfuscation to some literalists but they are actually just as relevent as those horn clause and continuation comparisons. I was surprised that Djikstra didn't mention a few other states besides California! Received on Wed Jul 22 2009 - 23:46:59 CEST