PS Consultants - ideas & solutions

Is history bunk..?
January 2002

We’ve heard a lot about the comparison of the net phenomenon with railways, the motor industry and even the South Sea Bubble, but I don’t recall much analysis looking at the more recent events of early micro computing boom during the 70-80s to see what we might transfer to predict the next phase of net progress. With the gloss well and truly off internet mania, and the autobiography of its most high profile disaster (Boo Hoo) going on sale, now is a good time to see if there are parallels to be drawn.

One thing is obvious about the early micro age; although in 1965 Moore’s famous law said that CPU power would double every 12 months, it was the applications and software that ended up driving the market for hardware; and the hardware that restrained the ultimate capabilities of software.

Let’s review the landmarks in software. In the beginning there was word processing, which was  natural at the time, since it was the means by which programs themselves were being written. Prior to that, did we really poke holes in bits of cardboard that were sent in weekly batches for some distant mainframe to reject with a cursory rude note? Yes, we really did; and kids today have no idea etc. (unless they are Linux aficionados and eager to explore the delightful Vi).

The first business microprocessor-powered computers were focussed on word processing. Who recalls the Memorite, Wang or Diamond word processors..? WYSIWYG was a long way from being possible thanks to the crude display technologies that were only barely ahead of teletype terminals. The sheer cunningness of a cursor addressable terminal screen was something everyone gathered around to witness until well into the 80s, although the way ahead for directly addressable video memory had been apparent from the earliest Sinclair home computers. Everything in the early days was proprietary – software only ran on one family of hardware – and the term “compatible” had yet to be coined.

In the networking world, there is an obvious parallel: the work done with proprietary transport standards that came complete with the proprietary hardware implementations – products like Novell’s IPX, Arcnet, Transnet and Microsoft’s NetBios lived in a world of their own and were not interested in anything like interoperability with competitors’ solutions.

Memorite, Wang and Diamond word processors were “turn key” products where the software was dependent on custom hardware; but the grand daddy of “portable” word processing appeared in 1979: Micropro’s WordStar – an application that could be installed to cope with different cursor addressable screens, which meant that it ran on hardware from different manufacturers who were started to adhere to a portable standard, the CP/M operating system, which Digital Research had introduced in 1976. Hardware manufacturers were obliged to compete, and the competitive issue became how cost effectively each would run the “standard software”.

Portability liberated the software industry and the floodgates opened. Next up was Ashton Tate’s dBASE database authoring language, then the Supercalc spreadsheet. Until the spreadsheet, the notion of being able to do “What-if parallel mathematics” had not existed in the world of linear calculation. And whilst all manner of attempts have been made to refine the idea, today’s Excel would be instantly obvious to the user of the original VisiCalc. This was the first application that really could only have been devised with a computer in mind – it was not a simple electronic version of an existing idea.

A net parallel is again obvious: the adoption of the open TCP/IP provided a transportable standard that could be implemented on any hardware and any operating system. Commercial Ethernet adapters appeared in 1981 – and in the same way that CP/M was bad news for the likes of Wang, TCP/IP was bad news for those of proprietary bent like Novell; but great news for the rest of us. Suddenly all sorts of different networks could interoperate and single dial up connection was all that was required to connect to just about any computer network on the planet. With such universal application, there was no way back, and thank heavens TC/IP remains a public open standards, and is not in the control of any one commercial interest.

Back on the microcomputer timeline, Digital Research was king of the operating system castle, and apparently impregnable. CP/M had cut it’s teeth on 8-bit CPUs – notably powered by the Intel 8080 and the Zilog Z80 – then gave way to MSDOS on the 16 bit 8088 for a simple act of marketing when IBM introduced the original IBM PC in 1981. This was an ancient technology that was in many ways retrograde, but it was a standard that again set the software developers free from complex compatibility issue, and with the IBM PC, Microsoft took full advantage of Digital Research hubris, and snuck MSDOS past CP/M as the standard.

However, the net is not free from monopolists. Cisco Systems boasts that it’s switches and routers power 90% on the Internet. However, the operating system of the Cisco hardware is based closely on Unix, and there are many examples of Unix-powered routing and switching solutions that keep alive the possibility of maintaining a wholly open Internet.

The PC and net stories combine at the point at which software became the most important issue. Web browsing became the focal issue. Netscape had been the “CP/M” of browsing – first, and believed itself to be impregnable. But then once again, it was stalked and strangled by Microsoft. But the limiting factor of the net is not CPU cycles, it’s bandwidth, and Moore’s Law does not take into account luddite telecom businesses and easily mislead governments, so we are now in the relatively uncharted waters of economics.

So what next for the net? The bad news is that railways, the motor industry, and the PC business were once vibrant markets that enjoyed the diversity of thousands of creative participants – and all ended up being controlled by virtual cartels and monopolies. Maybe the same writing on the net wall was apparent early on if you wanted to look for it, why should there be more than one web site offering to sell books? Answer: if that one site really can do an OK job of it – globally – then there is no reason for another to exist.

Clearly a more interesting future for the net lies in doing what the spreadsheet did for micro hardware – namely doing something that is not a simple conversion of an existing idea. An exchange site like eBay fits that profile, because it creates and empowers a market of hundreds of thousands individual participants. Whilst all previous “advances” have tended to marginalize the individual, the net is the first that may actually distribute power back to the creative individual, so maybe it’s not surprising that we are only just starting to work out what that really means.