Official Report search · Parliament Workbench

Chamber

Meeting of the Parliament 16 March 2011

16 Mar 2011 · S3 · Meeting of the Parliament

Item of business

Certification of Death (Scotland) Bill

Stevenson, Stewart SNP Banff and Buchan Watch on SPTV

I have a few observations to make, some of which pick up points that others have made and some of which are new. Dr Ian McKee talked about the importance of death certification feeding into health care planning. That is correct, but we must not fail to take account of the need for death certification to feed into immediate response to possible epidemics. Professor Stewart Fleming did not make reference to that in his definition of the three aims of the certification process.

Ian McKee also talked—absolutely correctly—about folie à deux. It is worth saying that in aviation some 20 years ago we had precisely that situation in the cockpit, when a very senior captain would often not be told by a very junior but recently trained and high-quality first officer that they were getting it wrong. In designing the relationship between different players in the system, we must be aware of the influence that respect for experience and seniority has and must ensure that a junior person can point out freely and frankly to a more senior person that they are not up to the standard that is required. Folie à deux was killing people in aviation 20 years ago, but training has changed and it is not killing people now.

Last week, I had the very great pleasure to be in Giffnock synagogue to launch a Jewish education project on the internet. On that occasion, I received representations on the particular issues surrounding Jewish burial practice, which are equally applicable to people of the Muslim faith. It is important that we take account of the fact that those faiths use burial rather than cremation and make sure that we acknowledge that and preserve those traditions.

Rhoda Grant talked about testing. It is worth observing that testing has more limitations than one might imagine. About 30 years ago, IBM produced a computer that turned out to incorrectly multiply 10 by 10,000,008. Every other calculation appeared to be correct, but it was established that to use testing to see whether they were correct would require every model of that computer that had ever been produced to run through exhaustive tests for more than 1,000 years. It is important to get the design of the system correct.

We have heard some discussions about computers and I want to make some observations, of which members might or might not be aware, that indicate the need for some caution. For the registration of births, Registers of Scotland provides 200 characters for forenames and 50 characters for surnames. Approximately 19 per cent of current registrations are for people who have three or more forenames, so that issue is not insignificant because people have more complex names than they once had. Until a few months ago, I was refusing to take my parliamentary payslip—I was still taking the pay, of course—because my name was not right on the payslip. I am James Alexander Stewart Stevenson and the system provided for only two initials, thus omitting the initial that I use.

Joining computer systems together is often complex when we look at the metadata, to use the technical term, that are associated with information. I say that in the context of my genealogical researches on my great-grandfather who was a coal miner in Bannockburn. He first appears in the record in the 1841 census. The difficulty is that he is one of 328 Stevensons who were working in coal mining in Bannockburn in that year. Having the ability to distinguish names is very important indeed.

Equally, even if we impose rigorous standards for data collection and entry, there might be difficulties. When I worked in the Bank of Scotland, financial services legislation was introduced that required that we collect people’s dates of birth. Our tellers found themselves inhibited in asking ladies of a certain age what their date of birth was, but they had to put in a date so they just chose a random date. We ended up with something like 9 per cent of dates being 1 January. A further 2 per cent turned out to be the teller’s own birthday and, for a further small proportion, the teller simply entered that day’s date and discounted the number of years. Computer systems are not just mechanical systems; they have to interact with the human effects that often surround the collection of data.

If time permits, Presiding Officer—no; I see that you are signalling to me to wind up.