On 08/02/2011 07:26 AM, Peter Alexander wrote:
Here's the thing: Unless you're developing a D compiler, there's no
difference between a QoI issue and a language design issue. An
issue is an issue, whether the fix is to change the design or to
change the compiler.
In my opinion, there's very little point in discussing a language's
merit without discussing the tools. A design without an
implementation has no worth. As Steve Yegge likes to say, a
language's popularity is all about the tools.
Fair point. It's just good to to use the right terminology. Otherwise if
you say "bad design" that's forever. "Insufficient implementation" is
transitory.
What's more constructive? Judging a language based on the Digital
Mars Hypothetical Future D compiler, or based on DMD, GDC and LDC?
What's the point of saying that a language has no flaws, but there
are no implementations of that language?
Probably the worse mistake is saying that the language has a flaw
because its implementation has one.
Andrei
