D string interpolation

[bitwise]

On Sunday, 4 October 2020 at 14:04:04 UTC, Mike Parker wrote:
> On Sunday, 4 October 2020 at 13:16:53 UTC, Imperatorn wrote:
>> [...]
>
> [...] so it will require at least one more round of community 
> review before moving forward.
>
> DIP 1036:
> https://github.com/dlang/DIPs/blob/master/DIPs/DIP1036.md

I hope they change the syntax to favor readers of code, not 
writers. As with D's templates, reading code is made more 
difficult by needlessly adding syntactical variations for 
language features.

Example:

writefln(i"I ate $apples and ${%d}bananas totaling $(apples + 
bananas) fruit.");

> $apples

> $(apples + bananas)

If I was trying to quickly scan code to pick out variables, 
there's two different things I would have to look for.

Example:

> Foo!int

> Foo!(bool, int)

Same problem.

> Foo;

> Foo();

Same problem.

I'm not sure if the reasoning is to try and reduce noise in the 
"common" case, or to appease people that think not having to type 
() will actually help them get things done faster, but neither 
reason seems valid to me.

Also, why would the less significant part of the construct be in 
the curly braces?

> ${%d}bananas

Why not like this?

> i"{apples} bar"

> i"{apples + bananas} bar"

> i"{bananas:%d} bar"

> Bar!(int)

> Bar!(bool, int)

> Bar();

> Bar();

In all honesty, part of the reason I lost interest in D is 
because of the type of quirks listed above. I don't believe the 
intended benefit, whatever it may be, is worth the sacrifice of 
consistency.