> Clark Jones writes:
>
> > A thought has occurred to me: A possible solution would be to "migrate"
> > to C++ (not a humongous project, since a quick look through a "tar -tvzf"
> > of a source-tarball reveals that it's mostly in C) and then use C++'s
> > ability to "overload" the normal operators to, in essence, construct
> > "custom fixed-point data types".
> >
>
> It's a good thought, and the use of C++ operator overloading is
> convenient for this kind of thing, but there is a rather large
> problem.
>
Don't want to start a language flame war, but after listening to several
weeks of C++ propaganda, I am no longer able to remain silent.
In my experience (which includes writing a C++ parser and semantic
analyser in C++ --- I'm not exactly a C++ novice), C++ is a slippery
slope. You start out innocently enough defining a few operators to
improve notation, and maybe a machanism for automating
reference counting. After a while, you start using more and more of the
metaprogramming tools built into the system and it becomes more
and more difficult to know what your program is doing (and if it is
doing the right thing). It takes ths utmost discipline to refrain
from this behaviour in C++, a discipline that is alien to the gung
ho style with whoch the C++ language was invented. In fact, the discipline
needed is very close to restricting oneself to the subset of C++ that is
essentially equivaleng to C semantically, which poses the question,
why not stick with C?
-- hendrik.
--
Gnucash Developer's List
To unsubscribe send empty email to: [EMAIL PROTECTED]