[Cuis-dev] [RFC] Modules/Namespaces for Cuis

[ken.dickey at whidbey.com]

On 2020-10-02 04:48, Gerald Klix wrote:
> On 2020-10-01 22:23, ken.dickey--- via Cuis-dev wrote:
>> On 2020-10-01 12:17, Gerald Klix via Cuis-dev wrote:
>> ..
>>> .. should we treat classes
>>> imported from other environments as constants?
>> 
>> How would one then add/update code to/in environments?
> Obviously I didn't make myself understood.
> I thought about a constant binding, with
> operational semantics like classes.
> `Smalltalk at:` is called when the class is compiled, when the code is
> run, the VM just
> accesses the literal in the compiled method.
> (Hence all the #elementsForwardIdentityTo:-voodoo in 
> ClassBuilder>>#update:to:).
> 
> I am sure, you -- being a proficient
> Scheme programmer -- know the concept:

Yep.  Brings back disscusions between Scheme and Smalltalk groups a 
Tektronix Research Lab in the 80's.  Note Allen Wirfs-Brock's proposal 
for a Modular Smalltalk (ref in http://www.educery.com/papers/modules/)

Note "A First Implementation of Modular Smalltalk (1993)":
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.56.6440

Note also ModTalk: https://www.modtalk.org/

Very different semantics and somewhat at odds with Cuis liveness goals.

Cuis also lacks Traits.

Many ways to do modularity..

Cheers,
-KenD