24 May
2010
24 May
'10
2:28 p.m.
On Mon, May 24, 2010 at 5:24 AM, Chris Cannam
<cannam(a)all-day-breakfast.com>wrote;wrote:
Reading a language is (for most projects) more important than writing
it. You yourself took the jackdmp code (in C++) and ported it back to
good old C because it was written "from the wrong school of C++" and
you found C easier to work with. Jackdmp is not exactly weird code --
it's written rather like pre-1.5 Java -- but its C++ is just not the
same C++ as you use. Similarly, for someone like me who has used Qt
for many years, Boost has always seemed largely superfluous and the
language that for you "is C++" is for me something a little bit alien.
Is it possible to write C++ in such a way that every competent C++
developer is happy to work with the results without some sort of
re-education?
nice point. i think though that its always been somewhat invalid. two cases
in point:
1) i have on my bookshelf coplien's "advanced C++: programming styles and
idioms" which came out sometime in the early/mid 1990's. it looked alien to
me when i first saw it then, and picking it up again, it looks pretty alien
to me now too. it represents a brief snapshot in time of what "good C++" was
supposed to look like. it pushed the envelope initially (i think that
coplien was the originator of the pimpl idiom, for example), but at this
point it looks pretty dated.
2) gobject... written in C ... as has been noted in this thread already,
despite being "analysable", its a style of C that some/many people find very
difficult to work with. so much so that its ended up being wrapped, replaced
and so forth. so the issues here are not restricted to the many headed hydra
called C++