I attempted to gather the costs of daylight saving time here. Software costs -- particularly in 24x7 data centres -- are significant. Another interesting angle is in this paper, looking at effects on equity markets.

The Barclays incident referred to in my post apparently turned out not to be time-change related, but they obviously spent time thinking it was. In my experience, the 1st Jan 2000 caused fewer IT problems than the average clock-change weekend.