[squeak-dev] Advice on new projects

> On Sep 20, 2009, at 7:47 PM, Michael van der Gulik wrote:
>> On Fri, Sep 18, 2009 at 6:24 AM, Ross Boylan <[hidden email]>
>>> Is there any namespace or similar mechanism that it would be advisable
>>> to use?
>>
>> Not really. People usually prefix their class names to avoid naming
>> collisions.
>>
>> I'm developing a namespaces architecture that will run on Squeak. It
>> works and I'm writing namespaced code using it, but it is not ready
>> for general use. If people are interested, I could concentrate my
>> efforts on making it usable, but so far I've seen more disinterest
>> than interest.
>>
>> http://gulik.pbworks.com/Namespaces - including links to an image.
>
> I think this work is important for more wide-spread acceptance of Squeak
> Smalltalk and its forks. I know it is only anecdotal evidence, but one
> of the first things I missed when starting an actual project was
> namespaces. After using languages that have support for namespaces,
> using the widely accepted practice of "just prefixing your classes with
> a two letter prefix" seems quaint and archaic. While that practice gets
> the job done, it's a turn-off to me, and I imagine that others coming to
> the language feel the same.
>
> So while you may be getting disinterest from the established community,
> again, I think that it would boost the popularity and acceptance by
> experienced programmers coming to Smalltalk from other 'namespaced'
> languages.

> Hi!
>
> Daniel P Zepeda wrote:
>>
>> On Sep 20, 2009, at 7:47 PM, Michael van der Gulik wrote:
>>>
>>> On Fri, Sep 18, 2009 at 6:24 AM, Ross Boylan <[hidden email]>
>>>>
>>>> Is there any namespace or similar mechanism that it would be advisable
>>>> to use?
>>>
>>> Not really. People usually prefix their class names to avoid naming
>>> collisions.
>>>
>>> I'm developing a namespaces architecture that will run on Squeak. It
>>> works and I'm writing namespaced code using it, but it is not ready for
>>> general use. If people are interested, I could concentrate my efforts on
>>> making it usable, but so far I've seen more disinterest than interest.
>>>
>>> http://gulik.pbworks.com/Namespaces - including links to an image.
>>
>> I think this work is important for more wide-spread acceptance of Squeak
>> Smalltalk and its forks. I know it is only anecdotal evidence, but one of
>> the first things I missed when starting an actual project was namespaces.
>> After using languages that have support for namespaces, using the widely
>> accepted practice of "just prefixing your classes with a two letter prefix"
>> seems quaint and archaic. While that practice gets the job done, it's a
>> turn-off to me, and I imagine that others coming to the language feel the
>> same.
>>
>> So while you may be getting disinterest from the established community,
>> again, I think that it would boost the popularity and acceptance by
>> experienced programmers coming to Smalltalk from other 'namespaced'
>> languages.
>
> Just a few notes:
>
> - The "established community" has a different view on namespaces in the
> sense that most of us value the absence of a lot of "pain" that the other
> "namespaced languages" thrust upon us.
>
> - I too :) have implemented a namespace proposal for Squeak, perhaps the
> best article I wrote about it is here:
>
>        http://swiki.krampe.se/gohu/32
>
> It is written in a cheeky style because as you may guess, the subject is
> slightly touchy :). I wrote that stuff because it was a fun experiment.
>
> And oh, I did fire up my code in latest trunk when I was down in Brest, and
> it seemed to work fine.
>
> Regarding the "boost" you mention... well, perhaps. But it is hard to make a
> really good non-intrusive namespace system that doesn't sabotage the rest of
> the Smalltalk experience.
>
> regards, Göran
>
> PS. No, I don't really think we NEED Namespaces that much (compared to other
> things). But if we do end up adding them I really, really, REALLY think we
> should do them in a lightweight smart manner and not just blindly copy some
> approach from another language with drastically different characteristics.
>
>
>