On Sun, Aug 27, 2017 at 10:56:49AM +0200, Hadrien Lacour wrote: > The point of using a compiled language is to avoid useless dependencies, even > if performances also count. > To be honest, it'd be more acceptable if it didn't rely on the most bloated > shell ever (baring fish, maybe). POSIX sh isn't that hard.
I add my 2c: Moreover, if not blind, you can see that everybody wants its "interpreted" language and wants to force it on all OSes out there. >From this perspective, the sickiest areas are: SDKs and server infrastucture components. You will have: perl, python2 _and_ python3, ruby, javascript, lua, php, swift, guile, bash not sh, etc. And for _each_ interpreted language, you add the bazillions of different and independant frameworks. Mecanically, that increases the 'technical' "cost" of reaching a "working OS". And the little added comfort provided by those interpreted languages is actually a big loss due to their unreasonable count and implementation "cost". But, the "compiled" language realm is turning to a turd too. It seems there are networks of insane/gangster people hard pushing with a lot of money (aka economic comfort zone for _selected_ coders) "compiled" languages with highly complex syntaxes: the main push is happening on c++ to phase out simple C totally (C syntax is not perfect and way too rich already but by far the best compromise). I don't even start on java/jee/sql, heresy againsty sanity. The worst components I have seen so far: - llvm - harfbuzz - gcc turning to a c++ kludge (got into the c++ "turdification" of gcc/cpp, yes the guys doing that are broken brains). Of course, you have file formats too. Look at PDF 2.0... 1/4 is "good" 3/4 is trash. Overall this is a really bad piece of work and is _very_ dangerous... as now it requires a huge and complex set of components to "make it work" (if I recall well, PDF 2.0 crapping was done mainly by microsoft people, them again). I can hear microsoft and apple laughting at us, and no, interpreted languages are in no way suckless. -- Sylvain
