Go: A new system programing language

[Michael Farnsworth]

On 11/10/2009 09:02 PM, dsimcha wrote:
> == Quote from Phil Deets (pjdeets2 at gmail.com)'s article
>> On Tue, 10 Nov 2009 21:21:27 -0500, Knud Soerensen
>> <4tuu4k002 at sneakemail.com>  wrote:
>>> Google have made a new language.
>>>
>>> See http://www.youtube.com/watch?v=rKnDgT73v8s
>>>
>> I watched the video. The language sounds like a cross between Smalltalk
>> and C, but with better concurrency support. I was somewhat underwhelmed,
>> but I do think the concurrency features are interesting.
>
> I watched part of the video, though I'm curious enough about it that I'll probably
> watch the rest later, esp. if Go keeps coming up around here.  However, for me
> personally "normal" static typing is too rigid for just about anything.  I would
> never choose a language that didn't either have duck typing or good templates that
> basically amount to compile-time duck typing.  I'm not sure how Go addresses this.

See http://golang.org/doc/go_lang_faq.html#inheritance

"Rather than requiring the programmer to declare ahead of time that two 
types are related, in Go a type automatically satisfies any interface 
that specifies a subset of its methods. Besides reducing the 
bookkeeping, this approach has real advantages. Types can satisfy many 
interfaces at once, without the complexities of traditional multiple 
inheritance."

That, my friend, is duck typing.

Incidentally, with pure duck typing I tend to feel a little bit...naked, 
unconstrained.  So while it's very flexible and allows any compatible 
types to "just work," it also makes it a little hard to define any 
useful constraints in practice.  Smalltalk makes this a tiny bit better 
because of its message passing syntax (the keyword arguments make things 
a little more clear), but I think D strikes a nice balance with 
templates and metaprogramming techniques, while still having some 
type-constrained relationships where they make sense.

-Mike