CP0956 – Resolution of Inconsistency Over Presentation of NGC weekly forecasts

Formal title: Resolution of Inconsistency Over Presentation of NGC weekly forecasts

Summary

NGC weekly (2-52 week) forecasts are received from NGC with the data identified by week number. This week number is converted into a week starting date for storage in the BMRA database. The week starting date is included with the week number for transmission in the relevant TIBCO message, and is also used to recalculate the week number for presentation on the Low Grade web site. It has recently emerged that whereas NGC use a convention in which the week starts on a Monday, the BMRA uses Sunday as the start of the week. As a result, the week starting dates being sent out in the TIBCO messages are incorrect (one day earlier than they should be). This is already leading to confusion over the week starting dates in the forecast data being sent out by TIBCO, as the BMRA expects the week starting date for week 1 of 2004 to be 4 January 2004 – whereas NGC’s week starting date for that week is 29 December 03 (because of the rules by which NGC decide whether there should be a week 53). The problem will become more widely visible, and potentially more acute, in December 03 when NGC begin sending forecast data for their week 53 of 2004 – which the BMRA will present as Week 1 of 2005. This may cause a number of problems, and it is not actually certain that the BMRA will be able to load the data for NGC’s Week 1 of 2005 when it arrives. It is suspected that the loader will fail because it will recognise the latter set of data as a duplicate.