Towards a widely accepted Smalltalk standard
Add your comments below about issues that exist with the ANSI Smalltalk standard. The hope is that we will be able to converge towards a Smalltalk standard that is really useful, and will be a guideline for Smalltalk implementors.Peter van Rooijen
- Fraction: ANSI says 3/4 better is (3/4)
Somebody added the following, but they are answers to a different question. The question is "what new things should go into the standard".
- Modularity: The minimum core set of classes and methods for a modular Smalltalk
- Syntax & Grammar: Typing, Comments, String Escapes, Variable Declarations,...
- Team Development: Headless repository services, OODB Services, Distributed Smalltalk
- Interchange: XML and other rich extensible forms
- FFI: Foreign functions declaration, invocation, marshalling, ...
- Interfaces: What do they mean for Smalltalk
- Frameworks: Graphics, Collections, Numerics, Streams, Text/String/Character, Networking,... what represents a good foundation and for what types of software are relevant
I deeply object to everything from #3 through #8 (and maybe even #9!). Each and every one of these requres the invention of things that do not currently have jack squat to do with Smalltalk. Standards should not be about inventing new things! Particularly things that just have nothing to do with Smalltalk.
These items have everything to do with
These items can and should be supported by the Camp Smalltalk world, if there is interest to pursue them. But, as topics for the Next or Future or Alternate Smalltalk Standard? No. Samuel S. Shuster
- Style (Commenting, Variable Declarations)
- Invention of non Smalltalk Intrinsic behavior (Headless Repository, OODB Services, Distributed, FFI, Marshalling, Interfaces, Networking)
- just plain You're Not Going To Need It extensions to existing behavior (String Escapes, ALL of the Framework sugestions except Networking)
- and last but not least, but most sad, the most gut-wrenching noxious affront to all that Smalltalk stands for... Typing!
Calm down, Sam. The current standard describes an exchange format that nobody uses. There is nothing wrong with coming up with one based on XML that people would really use. People have invented the team development things in Smalltalk 20 times or so, it makes sense to finally standardize on them. And most of the issues on the frameworks list are core Smalltalk. It is a mistake to try to standardize something that hasn't been done before, but just about everything has been done in Smalltalk, and when something has been done a lot and is fairly well known then it is ready to be standardized.
But the real problem with this list is that it just doesn't answer the question. The question was not "What topics should go into the next standard" but "what is wrong with the current standard". None of those describe the standard. Ralph Johnson
Is there any undertaking to update/upgrade the standard that's out there now (January 2005)? John Pfersich