The Development of Merlin

Eleven years is a long time for many things, including for a computer project! As the following timeline shows, the idea for the Merlin project is even older than that.

The current project was started in 1992, as very little of the developments which started in 1984 and 1989 could be effectively used ( though, of course, the experience gained was very valuable ).

Project Timeline


the Project

Related Events


In a way, the starting point for Merlin was the August issue of Byte magazine about the Smalltalk-80 object oriented language. It showed that computers could be made simple ( and even fun ) to use - very different from the CP/M style computers that were popular back then.

IBM introduced the PC


The idea of a Smalltalk-like computer for the masses went from day dreaming to action when the Pegasus Project ( which developed a computer for children and schools ) was started.

A microkernel, message passing Unix project was started at LSI - USP ( project SINDE )

Apple introduced the Lisa, making Windows and Mice industry buzzwords


It became clear that using Pegasus as a scaled down version of the more ambitious goals was resulting in a unsatisfactory hybrid, the project was split in two ( Pegasus died out a little later ), with the Smalltalk computer renamed to Merlin

Several alternatives were considered for the hardware and a bytecode-less Smalltalk was designed. Later the Smalltalk method names were translated to Portuguese and the "blue book" implementation was translated to C.

The Brazilian Congress transformed the "Reserved Market Policy" into a law by an almost unanimous vote

Apple introduced the Macintosh, the "computer for the rest of us"


The Merlin I prototype was designed and built

Digitalk introduced Methods, the first commercial Smalltalk ( for the PC )


"Inova Tecnologia e Informatica" was founded in association with Softec ( a leading PC maker in Brazil ) to develop Merlin

The Merlin II prototype was built in record time


Merlin OS, a very simple operating system, was written to help debug Smalltalk on the Merlin II

The Merlin III was designed and built. It was a dual bus 68020 machine - it had both NuBus and 8 bit PC slots

First a QNX port to Merlin III was canceled as well as the Smalltalk effort, being replaced by new OS design written in an Objective-C like language. Then the hardware project was scrapped, and the software was moved to the PC AT

Xerox spin-off Parcplace was created to finally do something about Smalltalk

"Self: The Power of Simplicity" paper was published at OOPSLA


The association with Softec came to an end


The project was restarted at LSI - USP.


The detailed design of Merlin IV was finished

Self 1.0 was released at Stanford


A 64 node machine was built at LSI ( MS8702 )

The Self project moved to Sun


The project was restarted once again. This time as a "home-grown computer"

The "Reserved Market Policy" came to a eagerly awaited end

Self 2.0 was released


The Merlin IV Prototype runs its first programs

the first lines of tinySelf in C were written on a DOS machine

Self 3.0 was released


The first Merlin IV prototype stopped working ( the stress to the board of constantly replacing EPROMs proved to be a problem ). A second prototype was built

Merlin's Web pages at LSI - USP let the world know about the project


Work on tinySelf was restarted, this time on Linux

Self 4.0 was released

I would like to thank the people who have helped Merlin in the past:

see also:
| new3 | | intro | | faq | | history | | runtime | | tutorial | | tiny | | plan |
back to:
| merlin | | LSI | | USP |

please send comments to (Jecel Mattos de Assumpcao Jr), who changed this page on Sep 1, 12:09 .