by Spike » Sat Jun 28, 2014 2:01 pm
C often has lots of dependancies, lots of paths, lots of different tools that all need to be installed and configured separately.
Then you have windows users using different versions of those tools (msvc2003,2005,2008,2010 etc), different profiles (enterprise, professional, express), or completely different toolchains (msvc, mingw, cygwin, dev-c++, eclipse - yeah, okay, most of these use mingw but they're all configured differently) etc.
And C++ generally has far more dependancies, from what I've seen.
at least on linux, its just: apt-get install dep1 dep2 dep 3 && ./configure && make install
just make sure your .so symbols don't get merged with other symbols in a really awkward way...
QuakeC is self contained. There's no libraries, there's no dependancies, the compiler used is small enough to just be thrown into the source zip. Double-click the compiler, then run your engine exe. done.
That said, perhaps people find QC so simple because the base code is fairly small yet sufficiently complete to produce something worthwhile, such that its easy to find all the right places to hack it to add other things too?
Writing QC from scratch is a pain, as much as writing java from scratch can be.
With quake, you already have the rendering code provided by the engine, just spawn an entity with the right model and you can instantly see the results.
imho, quakec's convienience comes from the fact that the compilers are small and self contained, rather than the language itself.
.