Alan Lovejoy wrote on Mon, 29 Aug 2005 15:15:08 -0700
ShiningRay @ Nirvana> "But I haven't seen many actions of SELF development.
I cannot find any application written in SELF. why?"
Several applications have been developed in Self, but all of them part of someones's PhD thesis. Most aren't available for download and so aren't particularly visible on the net (in no particular order):
Ole Agesen: Mango (A Parser Generator for Self) Albertina Lourenci: EcoDesign (CAD for architecture with tiling) James Noble: Tarraingím (A Program Animation Environment) Ivan Moore: Guru (refactoring system - http://selfguru.sourceforge.net/) Reinaldo Silveira: SelfHDL (simulation system for integrated circuits)
That two of the associated texts are in Portuguese certainly doesn't help with the visibility issue. Sorry if I missed anyone.
Might as well ask why C++, Java, Visual Basic, Perl and Python are generally used in preference to Smalltalk by most development projects.
Napoleon (among others) said "don't attribute to malice what can be explained by incompetence". While Smalltalk suffered greatly at Digitalk-ParcPlace due to incompetence over the Java hype machine, Self had to endure malice within Sun. I can understand a company wanting to be 100% behind a given technology and cutting back on the rest (think Apple and the Newton for an example) and the core Self team was actually allowed to linger on for a couple of years before being relocated to the Java group. But on the marketing side I was very surprised at the hostile reaction of some Sun people at the mention of Self (well, after explaining what it was since they hadn't heard of it before).
Only after 2000 did things become more reasonable at Sun and development of Self and variations was restarted and now you have things like Fortress. They are no longer counting on a "one language to rule them all" future.
One issue with Self is the lack of either a commercial IDE with major vendor support or a Squeak-like open source development project for the language and IDE. Self has always resembled more a research project than a serious attempt at a production-ready development platform.
I don't understand - Sun wasn't able to utterly kill Self in 1995 because they *had* previously released it all as open source. A major problem was that you needed a Sparc workstation to use it. I bought an Ultra 5 in 1998 just to run it but am probably the only one in the world who would do that. Several attempts to port it to x86 Linux in the 1990s didn't get very far as the VM was a very large C++ program. Squeak's design was in part a reaction to these problems. Only in 2000 did Sun decide to release a Mac port that had been done quite a while before and this refactored code finally did get ported to the PC a little later. The Sparc is still the best machine on which to run Self, however.
Self has its share of bugs but no more than Squeak, in my experience. And while StrongTalk was supposed to bring Self's technologies to a production quality system I don't think they did produce a vast improvement.
-- Jecel
squeak-dev@lists.squeakfoundation.org