Sharing in D

[Sean Kelly]

== Quote from Steven Schveighoffer (schveiguy at yahoo.com)'s article
> "JAnderson" wrote
> > My option is that the experimental branch of D should be correct first and
> > backwoods compatible second.
> That is a great recipe for the forking of D :)  I'm not saying that breaking
> changes should be unacceptable, I'm just saying that there better be a
> really compelling reason to break things.

libd forked the language and is far better for it.  Tango has avoided this
route because our primary goal is to provide a library for D.  This seems
to have turned out to be "provide a library for D 1.0" however, given
the lack of support for certain programming methodologies in D 2.0 that
Tango uses pervasively, as you're no doubt aware :-)  This leaves Tango in
a bit of a pickle, since it means either changing Tango or forking D if we
ever want to move past D 1.0.

> I believe Walter's view of "wild west" programming is completely off target.
> Many good multithreadded applications exist today, without the aid of this
> shared/unshared view.  I don't see how having this is going to magically
> solve any threadding issues.  If anything it's going to just pop up as an
> annoying obstacle that is frequently bypassed with casts, making the system
> useless anyways.  I don't know if this can be shown until the idea moves
> from theory to practice, so maybe we'll just have to wait and see.  This
> could be a stepping stone to a really great system, just like the first
> const incarnations were awful, but the one we have now (save a few missing
> pieces) is pretty good.

Systems programming is the low-down dirtiest part of the Wild West, too.  It's
a realm where programmers do all sorts of things that a general applications
programmer would be canned for, because it's necessary and because they *do*
know what they're doing.  I am very much concerned that the "future of D" is
as a general applications language and that its systems language roots will
be left to rot.  And who needs another applications language, anyway?  The
world has a million of them.


Sean