Letter: September bug

Sir: Your forecast of additional millennium bug problems after 9 September 1999 (report, 19 July) is very unlikely to be realised. Speaking as a now geriatric former software writer I am 99.99 per cent certain that nothing of note will happen on or around that date.

The basis for your original concern surrounds the practice, when using fields of fixed width, to allow invalid values to represent the end of a sequence. In the 1970s, dates would have been of the form "ddmmyy", resulting in the familiar millennium bug problem: 1999 was of course not invalid, but sufficiently far in the future to convince most developers that their work could not possibly be in use by that date. However, the "end of file" sequence would be 999999, or the 99th day of the 99th month, 1999, not, as your article suggests, 090999.

WILLIAM HUDSON

Abingdon, Oxfordshire