Table of Content  Branch Up  

Design Phylosopy

A simple computer designed to taste old and serve well.

"What we do today, tomorrow sounds old" --pointed out Ken Olsen in 1982 referring to Digital's effort to enter the personal computer market. I would add that obsolesce is not a bad thing for technologies that reached their momentum of glory at some point in history, and that's exactly the case of "minicomputers" back in the 1960s and 70s.

Most people today recall their first encounters with early personal computers unaware -in most cases- that computers started to be popular much before than that, possibly when they were attending elementary school. What actually introduced the computer to the popular culture was not the so called "personal computer" but those refrigerator-size monsters that back then were called "minicomputers".

Heritage/1 has been designed around that concept. It is not a "mainframe" because mainframes are designed to carry out heavy work loads and their design puts an emphasis on reliability (to which end they employ generous amounts of redundancy). It is not a "personal computer" either because it lacks a video monitor, a keyboard and a mouse.

Heritage/1 is a minicomputer and she has been designed to work exactly in that "classical" fashion. Any temptation to modern thought has been intentionally removed from the mind of the author. The modern concept of "fixed disk" is an example: back then, data (and software) did not reside "into the computer" (as part of it) but "outside the computer" supported by removable media such as tapes and removable disks.

It is hard for modern people to put the PC experience aside in order to get into the minicomputers spirit. But even harder is to conceive a computational environment which is not interactive at all. Not being able to run a program by just entering its name as a command. Not entering data from the keyboard directly into the application's interface nor asking short question for getting immediate answers.

Computers, however, worked in "batch processing" for more than twenty years: from early 50s to long after UNIX were invented in 1972. Software and data were entered from files and results were obtained thought the printer after minutes (if not hours) of silent computing.

Today is hard to imagine such a work flow, specially while designing a homebrew minicomputer. It seems to be no excitement in writing programs that just read files and make printouts.

But that's precisely why I encounter "batch processing" to be a fascinating field for exploration. I consider it a mental challenge because it will require for me to rethink Computing all together up to the point that my current experience as a computer user is not use.

Time-Sharing (multi-user) is where the excitement becomes obvious. I suspect, however, that it's going to be extremely hard for me to produce a Time-Sharing system for Heritage/1. That is why I don't plan to attempt such a thing at present, not before I've gained some experience with system development in general.

The first idea behind Heritage/1 is to be a minicomputer. It has to do with punch and magnetic tapes, teletypes and video terminals. Culture can not be explained (otherwise it was Math) so I won't try to.

Some other ideas are the following:

1.- Meant to be useful (as minicomputers were at their time)
2.- Reasonable performance (no annoying waits)
3.- Simple architecture (no fancy features such as Virtual Memory)
4.- Data and software are "external" to the computer (no ROM, no Fixed Disk)