On Sunday, 8 September 2013 at 06:47:14 UTC, Walter Bright wrote:
On 9/7/2013 11:08 PM, Peter Williams wrote:
In summary, you've gotten rid of the need for this type of
duplication so why
would you introduce it?
I believe that is covered in the "Rationale" section of the dip.
IMHO the rationale of the proposal il pretty weak:
- You can't have a 1:1 correspondence with translated C++ code,
so the translation barrier can be lower.
- You can't read _easily_ the code.
The first is not a problem, if it is true that D avoidance of
duplication is better than C++ way of doing that stuff (and
that's a C++ problem, as Peter suggested).
I would also add that I don't think at all that this is a
concrete translation barrier: usually I start copying and pasting
the C++ header in the D code, and then filling the methods
translating from the cpp part one after another.
The second point is more subtle, as we are talking about an easy
_navigation_ in the code in the editor, we are talking about
being able to "gain a sense of familiarity" with foreign code?
The former is something that should not impact over the language
at all (alas, C++ navigation, back and forth between header and
implementation is a mess),
The latter resolved by D with DDOC, which it is perfectible BUT
is _today_ a wonderful tool for strangers: the D library section
on DLang site is there to prove it. You have at a glance all the
definitions, documented and in sync with the last compilation.
What is missing from that?
- Paolo Invernizzi
