Re: Linux' (and other OS's) code patterns present in device drivers, the kernel and userland code ...
> All programming languages have "the language" with its grammar, operators, oddities which are a result of trying to shoe-horn too much into an "elegant" grammar only to see springs start to poke through the fabric.
~
yet, since all languages are syntactic anyway (some sorts of protocol
handlers) and all syntactic games (like chess) are optimally closed,
so in those cases "reinventing the wheel" makes even less sense.
~
In case of lacking language features in B, you can always roll out
the inner obfuscated constructs from A and modularize them for another
language. Some of this is happening with embedded code for the so many
different platforms
~
There are cases which are not that obvious, say pointers in C but not
in Java, but the closest implementation you can reproduce in Java is
through the command object pattern
~
> Generally some languages have features others lack and for kernel programming you pick one that pairs features and performance, like C.
~
Well, C is here to stay. As part of the beauty of languages (even
syntactic ones) as tools their greatest feature! is that you use
languages to speak about, improve and design other languages
~
> Unfortunately with great power comes the ability to create something that only the original author with his particular set of development approaches, can effectively work with.
~
well, there will always be "poets" out there! But they will always be
about meaning and not "language" per se. I think
~
> So there's probably a few more years left for obfuscation.
~
;-)
~
I was thinking about such things from a different angle, but I think
I get your point
~
lbrtchx
Reply to: