My first testing role was in 1995 or -96. I had been headhunted into our organisation’s Quality Assurance Team during 1994 when we were engaged on a major project, but this was a regulatory price-setting exercise, not a software project. Software was involved, but that was in the hands of contractors (a Big Name accountancy firm); we only developed the in-house resources to do the software work after the contractors moved out and left us to fix the bugs in their software…
I started out working on data quality, liaising with independent civil engineering consultants to try to ensure that the numbers utility companies reported back to us were robust. After that, I examined the implications of bringing in what was then ISO 9000 to address quality across the organisation. Gradually, I also took up requirements gathering for the data reporting the utility companies we regulated would be required to deliver to us. Only once those requirements were all gathered, reviewed and their relationships properly mapped did we start developing a data capture system. This was effectively a spreadsheet; my testing consisted of working through the embedded calculations to confirm that they delivered the right answers - not as easy as it sounds, because there were a couple of calculations deep in the structure where the tech of the day could deliver different answers depending on whether you used a calculator or a computer! Look up the Pentium 4 floating point arithmetic bug, though the issue existed before then. And I had to give some very senior people confidence that the machine produced the “right answers”.
As we underwent mission creep and each year, colleagues asked for more information, the ability of the system to work effectively required greater levels of sophistication. By the time I left that role, 15 years later, it was a thin client database, with data being scraped off into a much larger database on our main servers which was then used for quite powerful financial modelling.
Perhaps the biggest difference is that my testing work consisted of regression testing the same application for each of those 15 years. You will understand that by then, I was pretty jaded with the whole thing. At the same time, the organisation had gone through changes and a new senior management had decided to drop that approach to data collection, and having sunk 15 years of my professional life into it, I felt rather affronted that everything I had done for the bulk of my time with the organisation was being trashed. I had to walk away from testing for a couple of years and do something completely different just to get my head back into a reasonable place.