On May 7th hundreds of millions of viewers around the world watched the exciting climax to the English Premier League football championship match between home team and clear favourites Manchester City and QPR who were fighting to avoid relegation. If Man City won they would be crowned champions. But at 90 minutes fulltime they were 1-2 goals down, and with 4 extra minutes to play. At this point, viewers of SingTel’s IPTV MioTV in Singapore were left on tenterhooks as their screens froze, which they had been doing all through the match. By the time they came back on, Manchester City had incredibly scored 2 extra goals and they were crowned champions.
SingTel were all mea culpa the next day and week, explaining to their customers, and to the regulator, that “switching volumes were five times higher than the previous peaks” (Straits Times 15th May). Network overload – that is a telecom problem, an IPTV problem, not an over-the-air (OTA) broadcast TV problem. Peak loading is a classic problem of all utilities, to provision capacity to meet a demand that occurs only intermittently. But a fixed-line telecom company traditionally operates to a 99.99% completion rate. Internet outside the telecom sphere is best effort, but within the telecom sphere it must play by the same rules. So in this case the technology was a problem only in the sense that this particular technology cannot cope as well as OTA broadcast. But the real cause of the problem was insufficient investment in network capacity.
Wherever the problem lies, in the potential of the technology to have a problem, or in the commercial conditions under which it is operated, the fact remains that new technologies have a downside. For example, when library catalogue systems changed to computerised databases they worked fine until the computer crashed or the power supply was interrupted. Without the card-based system as a backup, library users just had to wait, or come back another day. The card-system never failed, even if it was less efficient in many ways.
IT specialists will explain that the most common problem is with the way programmes are configured and databases constructed. Computers are not flexible – not just yet. Their functioning is blindingly logical and God help the customer or user who fails the logic. Technological determinists will argue that it is only a matter of time before the technology becomes really smart, but that misses the point. Humans are not all that smart in the way they scale, configure and use the technology, and therein lies the problem. College computer classes are not widely known for teaching that side of things; they are strictly for the techies.
The problem is not only about new technology in a hardware/software sense; it also applies to new technology systems. In the 1970s the Hong Kong Telephone Company had a system in place to give an early warning of a fall in demand and therefore a signal to cut back on capital investment in new exchange capacity. The system worked to perfection and gave the warning, but the management knew from experience that demand never fell and so overrode the warning. Within a year the company was forced to go to the government for a hand-out, and stricter regulation soon followed. Hubris? Stupidly? Short-sightedness? All the above and much more. It’s the human condition, and new technology will never change it; can it adapt to it? That would be new technology.
