[self-interest] An OO history

[Jecel Assumpcao Jr]

Jay,

thanks for the links. I know that the Simula folks get upset at being 
forgotten in OO histories, but Alan Kay most certainly didn't use 
Simula in early Smalltalk implementations (as claimed in the first 
link). He did study the Simula compiler while a student at Utah. And 
the first name for Smalltalk was "Simulation Logo" (also known as 
Slogo, which expressed Alan's opinion of how fast the implementation 
was likely to be), so there is obviously a link. But there were 
important things that were different, like message passing. A lot of 
people think that is just a fancy name for indirect subroutine calls, 
and miss the whole point.

The path to Lisp OO was interesting - Carl Hewitt created his Actor 
languages inspired by Smalltalk 72 and later attempts to prove lambda 
calculus could express the same things led to Scheme. Later OO Lisp's, 
however, seem not to noticed and generated wildly different object 
models that were reconciled in CLOS.An interesting product, from the 
Self viewpoint, was ObjectLogo. It was the first prototype based 
language that I am aware of, but their message passing syntax was too 
awkward.

ParcPlace was the third attempt by the Xerox people "to get Smalltalk 
out the door". The first was the portable Notetaker computer in 1978. 
Then the special Byte issue, books and tape releases (to Apple, DEC, 
Tektronix, HP and later to Berkeley) in 1980-81. ParcPlace was spun off 
in 1987.

Don't forget Apple Object Pascal, Objective-C and other efforts to make 
hybrid OO languages.

Self was created by Randy Smith of Xerox PARC and David Ungar of 
Stanford, with the first description published in OOPSLA87. The 
language was later implemented by Dave's students at Stanford, and the 
whole group moved to Sun (I think Self 2.0 was the first release there, 
but can't remember very well).

-- Jecel