>>>>> Jeffrey Law writes:
Jeff> I think everyone would be best served if they realized none of this is
Jeff> personal -- it's about technical decisions in an effort to improve GCC.
GCC development is far from perfect. The recent model generally
seems to be effective, although there is plenty of room for improvement.
I am not trying to discourage feedback and comments, but
complaining is easy -- throwing stones is easy. The question is: How are
we going to improve it and/or fix it?
There are many things that we may want to do, but what can we do?
And most importantly: Are the alternatives really better or just
look better from the outside?
This discussion has strayed from the specific dataflow topic, but
for that specific project (and other projects), I would encourage the GCC
community to get involved and improve the design and impleentation of the
feature.
Waiting for perfect is not a good strategy. Competition is good,
but GCC developers generally do not reject offers of assistance. A lot of
separate, duplicative projects may not get done, while compromising on a
single design with everyone contributing to the implementation has a
better chance of completion. In the long run, we get a single,
functional, complete, good but imperfect, and easier to maintain feature.
David
