Traits (http://www.iam.unibe.ch/~scg/Research/Traits/) are first class 
interfaces?  I understand we could have them in Squeak soon if more 
were able to help.

I think it is important to question the assertion: "Smalltalk would be 
harder to write because I would have to be consistent...".  It seems 
very possible that the additional complexity and power of Traits, 
namespaces, components, whatever can be supported by tools that make 
programming easier in every way except backward compatibility.  Not to 
say that writing such tools is easy.

Richard

On Wednesday, March 3, 2004, at 09:19 AM, 
squeak-dev-request at lists.squeakfoundation.org wrote:

> From: Trygve Reenskaug <trygver at ifi.uio.no>
> Subject: Re: Some comments (was Back to the issue... (was RE: Squeak
> 	coding style...))
> To: The general-purpose Squeak developers list
> 	<squeak-dev at lists.squeakfoundation.org>
> Message-ID: <5.2.0.9.2.20040303175407.00c46c98 at uio-pop.uio.no>
> Content-Type: text/plain; charset="us-ascii"; format=flowed
>
> Ale,
>
> I agree, but I feel that we should possibly go even further. Smalltalk
> defines a language, its defined words are class names and message
> selectors. The current practice makes every class into a separate 
> dialect
> that may or may not be similar to some other dialect.
>
> Smalltalk would be easier to read if we made interfaces into first 
> class
> citizens. A selector would then mean the same everywhere that 
> interface was
> used. Smalltalk would be harder to write because I would have to be
> consistent and to invent good names that preferably didn't conflict too
> badly with prior uses. And yes, it would be important to comment the
> selectors; methods would only be commented if there was something 
> special
> in the implementation.
>
> I feel interfaces will be essential in the context of my IS21c project 
> and
> probably also Jacaranda.
>
> --Trygve
>
>