Is Smalltalk ready to make a comeback?

[Nathan_7]

This recently published article seems to suggest so. However, since JavaScript is eating the world, I don't know that any "new" programming language today can gain traction. Am I wrong?

[JamesCherrill]

I was Principal Consultant with Easel/VMark after they acquired the Enfin Smalltalk implementation (now part of Object Studio) in the 1990's. The claims made for Smalltalk are largely true. It was my first intro to formal Object Oriented programming and I never looked back. It was truely ground-breaking, and highly productive (the Enfin toolset included round-trip UML modelling and relational database persistance).
Having said that, I can't see it coming back in any big way. The language is too different from any of the current popular languages (eg no control structures) so you can't transfer skills as easily as you can between any of the C-derived languages. It's also very unsuitable for any kind of rigorous development - yes it's great having everything checked at runtime rather than compile time, and being able to redefine basic system classes is great fun, but neither of these lead to well-documented well-tested code.

It has a glorious place in the history of computing, but that's where it belongs, IMHO.

[rubberman]

Like James, I also did a considerable amount of Smalltalk development in the early 1990's. We were in the process of deciding whether to use ST or C++ for the development of FACTORYworks, a major semiconductor factory MES. We did a lot of prototyping in Smalltalk to prove our concepts, but in the end went with C++ for the production software. Because ST is so well designed as an object-oriented language, it was not difficult to take those prototypes and implement them in C++, though we had a LOT of framework stuff to do that ST supported out-of-the-box, such as reference-counted garbage collection, inter-object and inter-process message passing, etc. I managed the design and development of that framework. In the end, we had 10 million lines of production code, not one delete statement, and zero memory leaks. We could also plug in a shared library and voila, instant support for another relational database!

I will always have a tremendous respect for the inventors of the language at Xerox Parc - a place that was home to many of the technical innovations that we now take for granted.