[tpm] Complexity and language

[arocker at Vex.Net]

>
> In other words, most algorithms (once understood) are SMOP
> ('Simple Matter Of Programming')
>

The question is, what language features do those algorithms need for
economical expression? We could build everything from simple bit-flipping
operations, but wouldn't want to.

> To me, the next level up is dealing with concurrency, parallelism,
> and time-synchronism.

Good point. I'd forgotten "real-time", and of course that's a major part
of  operating systems. At the very least, that brings a need for queue
operations, where Perl's list-handling is powerful. Object-orientation
probably begins to be useful here, especially if the events are
complicated.

> After that you get into the domain-specific functions/algorithms
> which is a separate issue that may or may not have 'complexity'.
>
Mere obscurity, dependence on particular values, order of processing, or
length don't make an algorithm complex, it's more the degree of
abstraction, the types of operation to be performed and the number and
depth of relationships involved, I suspect.