On 15.10.2016 16:52, Lew Pitcher wrote:
[...]
>
> The number of features a program has does not always correlate to the number
> of words used to describe those features in the manpage.
Sure. Now can you guess what the OP intended to express? (Too many features?
Too few features?[*] Too verbose a man page (absolutely)? Too verbose a man
page (relative to the number of features)? To hard to learn shells features
from the man page? Too hard to find topics in huge man pages?) If it's that
clear - to me it isn't - why don't you just clarify how to interpret the OP,
or, yet better, provide an answer to the OP's question.
I think it's quite pointless to assume any modern shell to be simplistic;
they are, after all, programming languages, and there are requirements and
user expectations, e.g. for performance, that can't be satisfied if the shell
relies on external tools for every simple task (like string manipulation).
>
> I've seen featurefull programs with exceedingly terse man pages where you were
> expected to use the program's -? or --help option to determine what features
> were available.
Or a terse --help information and a complete description in the man page (as
for example in the given context of bash, ksh, zsh). What does that tell us
WRT the OP's posting? (Not much, I'd say, but feel free to explain.)
> Conversely, I've seen featureless programs (with only -V
> and -? options) with disproportionately verbose man pages.
I wouldn't consider specifically a shell like bash as featureless, so I don't
see what insights that comment contributes to clarify the OP's intention.
Janis
[*] The original topic was “Do One Thing, And Do It Well”; to satisfy that
tools should have a complete coverage of their specific application domain.
The converse interpretation (to restrict the possibilities of a programming
language to handle "OS entities") would require quite a bit more effort [of
the OP] to suggest in what ways and to what degree any restriction of bash
(or shells generally) should be done, which trade-offs we could accept, etc.