... and flops we hyped

For every success in the world of information technology there has been more than one flop. Stephen Bell takes a look at some of the flops.

For every success in the world of information technology there has been more than one flop; a courageous idea that proved to have fatal flaws, a concept that was okay or even brilliant technologically, but proved to have no market, no appeal to the user – and, of course, the innovation that came just too late and was steamrollered by a bigger and better technology arriving just behind.

Few probably now even remember PCS – the Personal Communications System. Cheaper then cellphones – or rather cheaper than cellphones were at that time – it required the subscriber with a portable phone to find the nearest pole-mounted transmitter and stand under it to make a phone-call. The cut-price cellphone steamroller put paid to that one.

Network Computing – now there was an idea whose time had come in the mid 90s. Get all those complicated bloated applications off the desktop, save desk real-estate and the monumental effort of maintaining all those desktops day-to-day and upgrading them when they began to crawl. Network Computers would never slow down because their tasks were essentially simple – handle the graphics coming into the desktop and the keystrokes going out.

And it had Oracle supremo Larry Ellison behind it; how could it possibly flop?

It did, attracting only relatively few enthusiasts. Users, it seems, liked to have control on their desks and didn’t trust the network for reliability and speed.

We can hardly cop the total blame for hyping speech recognition. It is an idea that has been with us since the early days of computing. In every digital era we seem to have been told that one of these days, we were going to be able to speak to our computers, and would be freed forever from the drudgery of keyboards.

Again, the technology has been developed to the point where it accepts connected speech, eliminating the need to make those awkward… pauses…. between…. each…. word, and the high inaccuracy rate that bedevilled earlier efforts. But it seems most computer users are happy with the keyboard, and don’t really want co-workers overhearing everything they tell their PC.

Speech recognition has established a small role in the voice-telephone market, and a new standard has been developed to allow users to “talk” to websites. But time will tell whether this is a capability that anyone – other than people with a disability that prevents them operating a keyboard – would actually want.

The Apple Newton also tried to break new territory in the human-machine interface with handwiriting recognition added to portability in a PDA. The unreliable performance of its most hyped feature made it famous in all the wrong ways in the general media. "I can't make it," says a character in the syndicated "Doonesbury" comic strip. "According to my diary, I have an appointment with Boll Jecksom at Cafe Fwblob."

Locally, even an attempt to put the Newton before a mass audience by having an emerging Shortland Street character leave one behind in a cafe backfired. "What's that?" asked one character. "Oh, it's just a yuppie diary," another replied.

The Newton became a curiosity of history, like the pre-Macintosh Lisa windowed computer, perhaps, an idea whose time had not yet come.

Never mind, Apple, we're sure Shorters gets you a lot of iMac sales.

On the operating system side Unix in its various forms has maintained its place against the tide of Microsoft. But unified, truly standard Unix? – that was a hope hyped through its successive incarnations, but which never eventuated. First there was Posix, then (wouldn’t you know it?) a rival standard from a consortium known as Unix International sprang up. We studiously reported Posix/UI efforts to achieve common ground, but no standard Unix OS ever eventuated. Like many lesser standards efforts, it capitulated under the pressure of individual vendors trying to add that one unique point of distinction as a competitive weapon.

With the parallel rise of Unix and Windows went hopes for a reform of the computer market away from mainframe centralization to distributed processing power. Yes, it happened, but predictions of the Death of the Mainframe – I have that slogan on a T-shirt from a 1985 NZ Computer Society conference – were, to say the least, a little premature. Big iron may have changed its sales message, from big processing power to secure storage for massive databases, but the mainframe is still very much with us, particularly in our large corporates and new-style “computer bureaus” and their datcentres.

On the networking side, the ascendancy of Ethernet was challenged by IBM’s Token Ring. We covered the development, we looked at user examples of it, we read the white papers and listened to the arguments of those on both sides, as to whether a disciplined passing around of tokens to allow one to speak – it always reminded me of the boys passing the conch in William Golding’s Lord of the Flies – was superior in speed and/or reliability to Ethernet’s CSMA/CD. That means “all shout when you want, and if you can’t be heard do it again.” Surprisingly, it seems to work better than the disciplined approch.

Marketplaces – electronic networks enabling customers and providers to relate to one another and exchange goods and information – look right now, in any case, to be one of the most thoroughly hyped flops. Leaders of the concept like Ariba, Broadvision and Commerce One, began to look financially sick earlier this year. Ariba pulled out of New Zealand, and while local marketplace operation Onezone, using Ariba’s software, sounded a note of confidence at the time, it was gone by October.

TelstraSaturn’s joint marketplace venture with SAP likewise failed to attract customer interest.

Perhaps marketplaces will undergo a resurgence; perhaps it’s too early to condemn them as a failed idea. But things don’t look bright at the moment.

Business Process Re-engineering was all the rage in the early 90s; a grab-bag of philosophies and techniques which combined the thoughts of management gurus with a hefty dose of supposed IT smarts. On the IT side, it was one of the periodic resurgences of the philosophy that you could define a business with lines and boxes on a chart or (better) on a screen. I recall the head of a major export organisation grumbling to me: “These computer people and business process experts come in here asking me: ‘how do you make a decision?’ I say ‘I sit here and look out of the window at that lovely view of the Orongorongos and I rely on 30 years’ experience’. You can’t put that into a computer.”

Well, on the broader front, people are still trying. with concepts like “knowledge engineering.” That one we can’t quite consign to ‘flop’ territory (yet?) but BPR has vanished into the mists of history.

Reduced Instruction-Set Computing (Risc) is another one we can’t consign completely to the dustbin; it’s still grinding away there in the depths of some computers, an unremarked worker-bee. It somehow never became the saviour of continued increase in computer power and simplicity it was supposed to be when Hewlett-Packard introduced it across its range.

The idea, in case some readers have forgotten, is to use only a few simple machine instructions in place of the huge repertoire that has built up in most modern machines, derisively described as Cisc (Complex Instruction-Set Computing). With Risc, a complex operation is simply built up of a large number of simple instructions, simplifying the management of registers and core processes in the machine.

In the event, the superiority gap proved not so great, and Moore’s Law seems to have trundled on relentlessly without it.

We still occasionally hear of document management, as a specialised discipline, but the high end of it has been absorbed into the broader-ranging but still fuzzy concept of "knowledge management." And the basic end, the retrieval tags, check-in/check-out and version control, are increasingly well within the capabilities of standard office software (with an upper- or lower-case o). Another victim of the steamroller.

And what about office automation? Maybe already in 1986, some of us sensed it was a dead duck - again as a specialised form of endeavour with its own distinctive hardware and software. But it was still being touted in the mid 90s - and in our own publication.

Again, a case not so much of flopping but of of melting into something subtly different and larger - the whole word-processing, business information presentation, workflow, information sharing, collaboration sphere. It's all there in various compartments of the industry, going by various names, but has never quite become the same unified concept we thought it would be in the early 80s.

As for the "paperless office." Did Computerworld ever "hype" that? Maybe, in a moment or two of absent-mindedness, but surely we all appreciated in our heart of hearts that this one was as floppy as the paper that office information continues to be printed on.

The sesquidecade (well I’m using the word, even if no-one else is) also had its fair share of hyped and flopped IT projects. The Incis police computer system is the one that will most readily come to most readers’ minds. From a technological perspective it clearly suffered from a faith in the longevity of IBM’s OS/2 operating system, on which all the PC environments were originally based.

An independent inquiry into the project found this was not the only technological misjudgment. Object-oriented application development was overused and too much faith placed in the maturity of this still-developing discipline.

But it was fundamentally too large and monolithic a project, too long in the making.

Being used as the whipping boy for every Police shortcoming from lack of information to shortage of funds for vehicle fleet maintenance only hastened its inevitable demise.

Similar great hopes were held for the National Document and Information System, a joint development of the New Zealand and Australian National Libraries. The Australians formally pulled the plug on it in 1998, as costs spiralled.

Eventually its slow development ran into the Y2K problem, as it appeared the libraries would be saddled with a non-compliant system.

The Department of Justice, in 1992, advanced a novel way of funding IT development, with a planned registires system, initially to concentrate on birth, marriage and death records. Suppliers were offered the chance to design and build the system free of charge, in exchange for the right to on-sell the information.

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about AppleAribaBroadvisionCommerce OneDepartment of JusticeHewlett-Packard AustraliaIBM AustraliaMicrosoftOracleSAP AustraliaTelstraSaturn

Show Comments