LATERAL THINKING

Our ADSL went down again yesterday while we were in Rethymnon. It took me 4 hours to get through to the tech support at Forthnet (our ISP) who eventually promised to get onto OTE (the PTT) and get things fixed but not until we had once again checked all of my settings and had had the usual conversation about them not supporting my modem (a Belkin combined ethernet hub, 802.11g router and ADSL modem - don't support it? - they've never even heard of Belkin for Turing's sake).

Up until recently OTE have taken a couple of days to fix these problems that seem to occur whenever any of their technicians even looks at the local exchange (last time they had switched of the router that should have forwarded ADSL traffic!). However, today everything was back up and working properly by 10 am.

And so it seems that while the ADSL service itself is not becoming any more reliable they have at least worked out how to fix problems more quickly (just as long as it isn't actually the w/e that is). We must, I suppose, be grateful for any improvement not mind how lateral a solution it might be.

OTE's "process improvement" approach brought to mind other IT solutions equally as lateral and in particular one incident in the early days of Windows NT. The asset management company I was working with at the time had an impressive set of DEC Alpha servers running VMS (64 bit even then) that ran their entire global business and they were taking a punt on NT for a small system on the strength that Dave Cutler who had architected VMS was also responsible for the design of NT. The Alpha servers ran 24/365 with downtime required only to re-org the Oracle databases occasionally. The NT servers on the other hand were up and down like a whore's drawers and so the IT director had a serious word with the head of Microsoft in the UK.

To cut a long story ... the MS honcho returned some weeks later with the latest fix from Redmond which was duly installed on all the NT servers but reliability did not improve the NT servers were still falling over twice or three times a day. At their next meeting the IT director pointed this out to the MS honcho who smoothly asked him to look at the uptime figures. Lo, and behold uptime had improved!

Later, the IT director bearded his head of IT services and demanded an explanation. How could it be that the servers were still falling over with the same frequency and regularity but uptime was up? And, as you have probably guessed, the answer was terribly lateral: Microsoft's solution to the problem had been to reduce the time it took for NT to re-boot! Holy Ada Lovelace, Batman! How smart was that?