D's auto keyword

[retard]

Wed, 13 Jan 2010 20:22:02 +0100, Lutger wrote:

> On 01/13/2010 05:49 PM, Justin Johansson wrote:
>> Happy New Year 2010 Everybody.
>>
>> Having resumed C++ nationality for the last few months, I kind of miss
>> D's auto keyword.
>>
>> I am wondering, though, from an OO/polymorphism perspective, and UML
>> and sound software engineering perspective as well, what does D's auto
>> keyword buy you except (perhaps) laziness (in keystrokes)?
> 
> A lot:
> 
> No types to be seen here (aside from the literal 2) so you can: 1.
> rename the type whatever bar() returns without ripple effects through
> this code
> 2. substitute the type itself for another type that obeys the same
> structural interface.
> 
> In a way, it's like not using magic numbers only applied to types :) The
> benefits are not limited too but more pronounced in templated code.

The main benefit IMO is that you don't need to repeat things twice.

sadklaskdjlaj<aslfkjals, asdkjasldjw<asldjalsdm, asdjslk>> foo = new 
sadklaskdjlaj<aslfkjals, asdkjasldjw<asldjalsdm, asdjslk>>(); // no thanks

Repetition happens every time one uses advanced data structures like 
lists, hashes, tuples etc. In cases like

> 
> auto foo = bar(2);
> auto baz = grok(foo);

I would actually explicitly write to type some times to document things 
better.

> All ML derived languages do this much more extensively and they even
> don't need an auto keyword for it. C# has var and VB.NET has Dim which
> mean the same thing. The nice thing with those languages is that if you
> code in Visual Studio, you only have to hover above the variabele and
> the type pops up. Most IDE's have this btw, including descent, so that
> takes away the concern of having to remember the flow of types inferred.

A tru 1337 koder uses vim or emacs, both of which suck at semantic 
analysis :)