Consultant John Fisher acknowledges that the well-known French expression plus ça change plus c’est la meme chose (loosely, "the more things change the more they stay the same") is not entirely true of the IT, now ICT, industry. But his 40 years’ experience has taught him that some things endure, for example the eternal search for the methodology that will automatically generate so much code as to start putting programmers out of work.
Though methods have improved, that Holy Grail, promised at least 20 years ago, is still not even on the horizon.
In other respects, ICT practice swings on a pendulum, he says, the clearest example being the movement from mainframe and dumb terminal networks through shared client-server intelligence and the PC revolution and back to Citrix-type server-centric applications and a minimal, if rather better-looking desktop interface.
Other cyclic trends worry him more. One of the early slides in his presentation to the NZ Computer Society late last month epitomised “point solutions” by showing ancient Egyptians building the four sides of a pyramid independently with disastrous results. Database techniques ensured solutions were more integrated; but in the last part of his presentation, he remarked that “we’re still arguing about methodology” and that certain modern rapid development techniques seem to have forsaken enterprise architecture and returned to point solutions.
Fisher opened by saying he wasn’t going to present a “history of computing” nor a personal “potted CV”. His presentation did have aspects of the latter, which ranges from work on Burroughs adding machines programmed by setting pins in a tray through early IT development for British European Airways (the short-haul part of what later became British Airways), vendor ICL and its consultancy arm Dataskil, to New Zealand with Inland Revenue and Transpower among many others.
His IRD work allowed him to point out the underestimated effect of ICT on large-scale politics and economics. Data warehousing at the IRD had allowed the department to dispense with the annual individual tax return, he said, while the Customs system Casper in the early ‘80s was essential to the end of import licensing and a far wider range of imported goods for the New Zealand buyer.
Fisher’s “CV” offered a perspective on how far the industry has come in 40 years, or any subjective subsection of that period that members of the audience were familiar with. Several personal reminiscences emerged, and before the talk, the longer stayers could be heard outbidding one another on the skinny capacity of the machines that marked the start of their careers. “I worked on an 8Kb Elliott, but those were ‘words’ of about 12 bits each,” said one veteran. “Well my first machine may have been 12Kb, but they were six-bit characters,” another replied.
The “better-looking GUI” brought reminiscences too, of the early windowing systems on the Xerox Parc machines and the Apple Lisa. Several of the audience recalled confident predictions that the clumsy mouse would not last as a pointing device. The Parc machines had an alternative called the “cat” (yes, there was an acronym); a pad over which the user stroked a finger. That proved the more unpopular, but of course returned many years later (as a square rather than Xerox’s circular format) on laptops.
The desktop PC has undoubtedly brought major benefits, but Fisher points to the continuing risks of “PC anarchy” — the development of multiple inadequately-linked PC-based systems. When he was at Transpower, he says, there were 70 Access databases in the organisation; now they number more than 2,000.