On 11 okt 2014, at 07:11, Walter Bright <[email protected]> wrote:
> I've authored several DIPs, too. That aside, pragmatically most DIPs will not
> be implemented. New proposals for improvements fill the n.g. every day.
> Having a DIP that fails is still far better than implementing a feature and
> having that fail.
Of course they're not implemented. Most DIP's are ignored.
> DIPs, at a minimum, help prevent people from reinventing the wheel, they can
> take an existing DIP and update it.
>
> D is a fairly complex language, and there's a lot to learn. Shouldn't we be
> conservative about adding more?
Absolutely. But I can't really agree on the features that do get added and the
ones that don't.
Example: you don't think it's a good idea to support regular functions without
braces, something that bearophile has suggested several times. I like that idea
as well. But you thought it was a great idea to support eponymous templates
without braces.
> I don't know how we can add features to an already complex language at a
> rapid pace and maintain coherence. I don't know any language that has
> succeeded at that.
Right, but it's all about choosing the correct features to add, that can later
be used to build new library features on top. AST macros is a perfect example
of this, but also one of the most complex features to add.
--
/Jacob Carlborg
_______________________________________________
dmd-internals mailing list
[email protected]
http://lists.puremagic.com/mailman/listinfo/dmd-internals
