Testing, testing

This week's Computer Weekly reports on trials of 'chip and Pin' retail systems being run by shops and other retail outlets in the Northampton area. Chip and Pin allows debit and credit card transactions to be authenticated by cardholders inputing a personal identification number rather than by requiring a signature. Such systems have been used by many European banks for a considerable time (retailers often require independent ID too, such as a national identity card).

So what's my gripe? CW reports the implementation director for the project - covering 1,000 outlets and 200,000 cards - as saying: "The key aim of the trial is to educate users and gauge their reaction - the technology has already been thoroughly tested by retailers and their suppliers [my bold] before the trial went live."

However, when we read on, we find that: 'feedback from customers is already leading to subtle changes in the chip and Pin systems and processes operated by retailers"..."Consumer feedback is also helping retailers overcome some of the inevitable teething problems with the technology"...."some customers felt the Pin pads were not private enough, requesting a visor around the number keys"..."There have been some issues with the positioning of the keypads"..."The general expectations was that staff would put the card into the reader, meaning all the customer had to do was enter their Pin. However, feedback from Safeway and others has told us that consumers are keen to stay in control of the whole process"...

I'm a little amazed that machines have clearly gone into production and major assumptions made with such inadequate 'testing'. Time and money undoubtedly could have been saved. It's clear that there has been no user input at any stage before manufacture and pilot rollout of these systems, so how the project's managers can claim they had "tested the technology thoroughly" before this trial beats me. Or perhaps they believe that 'testing' means testing lines of code or somesuch.