I wouldn't spend much time on rewriting classic utilities in D. There's
plenty to be done anew. One category of tools to be written in D are
utilities aimed at D itself (parsers, analyzers, Thrift bindings,
protocol buffers bindings, code for DB interface, etc.)
Andrei
Indeed.
It could be reasonable to convert existing C or C++ code to D if you're going to
be heavily changing it, but converting pre-existing applications which are
currently in C or C++, and which you don't need to maintain, seems like a waste
of time. It _could_ be an interesting exercise in how to do things in D and
could very well show shortcomings in D, dmd's current implementation, and/or
shortcomings in Phobos, but then so would new applications.
At this point, if I can choose what language I'm going to write something in,
I'm almost certainly going to choose D (though obviously stuff like GUI apps may
not really be properly feasible in D yet, and some things are just gonig to work
better in other languages), but I have enough to do (and not enough time to do
it) without spending the time to rewrite entire, working applications in D.
- Jonathan M Davis
All true. I just figured that it's 1) a nice benchmark to show that
yes, D can replace C/C++ and 2) a way to review the existing utils and
potentially make them work a bit better. I also have a sneaking
suspicion that there may a few hidden bugs that could be prevented by
good D programming.
Casey
