% fortune -ae paul murphy

Everything that can be invented has been invented

If Charles Duell, in 1888 the Commissioner for the U.S. patent office, actually said that, then he would, of course, have been wrong; but I think the claim may have some applicability to the business PC in 2008.

The business PC keeps getting cheaper to buy but more expensive to own, and it keeps getting faster, fatter, and more complicated too - but what does it actually do today that its predecessors could not?

Today the PC runs a GUI based applications environment, it offers network connectivity, and it multi-tasks - but my Sun 160 did all of those things in the mid eighties. Not as quickly, of course, and without the common infrastructure like web browsers and the internet that didn't exist then - but email worked, users could retrieve documents from NASA and other sources using on-line text search from McDonnell Douglas and BRS, and there's really nothing obviously incompatible between what can be done on Solaris or a Mac now and what that machine could do then.

Notice that this is about differences in kind, not differences in scale or application - my 160 wouldn't play an mp3 file, but it could process audio; it couldn't store a 10GB file or run ZFS, but it could store files on disks and tapes - meaning, overall, that there are lots of things the PC does now that the Sun 160 could not do then, but the reason you can do them is that the infrastructure now exists - and because that has nothing to do with the PC itself, there's really no reason to believe that you couldn't dig a 160 out of that Unix museum frequent contributor TonyMCS loves to talk about and make it do those things.

At the time, of course, the 160 was a very expensive, very high end, desktop intended more to compete with the Vax then with the x86 world - and that fact makes it possible to argue that the mid eighties desktop Unix foreshadowed two very different computing futures.

The down scale future is one we got in the nineties: the high cost, low security, Walmart style, made in China quality, kiddie clicker focused, world dominated by the Microsoft PC - but the other one, the one built around extending Unix into the Plan9 world, around Postscript based network displays, and around high productivity, user focused applications like Oracle Office, was also there - and has neither materialized nor completely gone away since.

It's not highly visible, but it's there as a kind of underground movement: Solaris 10 (really SunOS 3.0) pushes in that direction, so do Sun Rays, Apple's iPhone, and OpenBSD.

Ever watch the Terminator movies? There's a fundamentally hopeful theme hidden in all the violence: the machines foment nuclear war, people are almost uniformly crushed by the machines, but a few continue fighting and everything we learn about the future suggests that the kid survives to defeat the machines - and thus re-establish human values, dignity, and progress.

So here's a frightening suggestion: look closely at the all Microsoft enterprise today and try to honestly answer two simple questions:

  1. what's all this stuff do that we couldn't do in the 1980s? and,

  2. what are the odds that the other path to user empowerment and distributed computing not only would have been better, but will eventually re-emerge, crush the machines, and declare thirty or forty years of the PC as much a multi-trillion dollar disaster as the failure to exploit nuclear power?


Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.