[ Jocelyn Ireson-Paine's Home Page | Publications | Dobbs Code Talk Index | Dobbs Blog Version ]

Three Men in a Bar Found a Spreadsheet Society

In Greenwich Market was a stall selling New Age kilts. They're black and velvet, and men wear them for night-clubbing. Apparently. I shan't buy one; "Taste of the Algarve"'s pies are more my fancy, filled with pumpkin, butternut, nutmeg, onion, and a smudge of Parmesan, with cinnamon for sweetness. I could drink raspberry beer in Royal Hill's Greenwich Union pub, run by a German braumeister. Or walk the Thames to Woolwich Royal Artillery Museum and learn how our army adopted khaki to avoid being shot; or further to desolate Crayford Marshes, where a moat surrounds ruins owned by William the Conqueror's half brother. They were inhabited right up to 1935. I might just stroll to the lovely clean architecture of the Royal Observatory's Flamsteed House to watch tourists giggle as they peer through the telescope in the Octagon Room. A notice near the telescope explains the discovery of Pluto; and for sure, as you look through it, Pluto is what you see. Stuck to the lens is Pluto the Disney dog.

I'm going to Greenwich next week for the 2008 annual European Spreadsheet Risks Interest Group Conference, where I'll be presenting my components work and learning what else is new. So I thought I'd interview one of the founders of EuSpRIG, Ray Butler, about how the group came to be.

Ray: I know vaguely that you, Dave Chadwick, and Pat Cleary founded EuSpRIG, but I didn't get involved with it until 2001, so I don't know the early history. What inspired you three to set it up?

I'm not sure if there was ever an inspirational moment, only a succession of happy accidents. But if inspiration there was, it came from Ray Panko, who was the catalyst.

Ray Panko, I should tell readers, runs the Spreadsheet Research web site. He's particularly interested in spreadsheet errors and how defects in cognition both cuase them and prevent us seeing them once cuased. This year, I see from the programme at www.eusprig.org, he is giving a talk on Taxonomy of Errors Revisited.

Last year, Ray told us how the broad picture our mind paints of the world is actually made of bits and pieces. Our eyes don't move steadily, but in jerks called saccades. Also, we are only aware of a tiny bit of what happens around us. The brain fixes this to give us a high-level view of what's going on; but in between the bits and pieces, the brain can make errors, and these remain inaccessible to our consciousness. Errors such as not noticing mistakes in long formulae. Anyway, can I return to the founding of EuSpRIG?

Going back to prehistory, I had been Customs's worrier-in-chief about spreadsheets since the 80s. The problem was businesses using spreadsheets for tax calculations and making the usual dog's breakfast of them.

That's the UK Customs and Excise, isn't it? Who collect Value Added Tax, which for our non-UK readers, I should explain is a sales tax. Is that right?

Close enough. Customs also collects duties on imported goods, the taxes on the nation's "Bad habits" — drink, tobacco, gambling, road fuel, etc. — and protects the country's borders against drug and other smuggling.

My interest was piqued in about 1985 or so by an article about spreadsheet errors in Computer Guardian, which pointed me at Spreadsheet Auditor.

Spreadsheet Auditor?

This was a stand-alone MS-DOS application that read a Lotus 1-2-3 .WKS file and allowed one to test for a number of common errors, find the trail through from entry data to end result, identify families of cloned cells, and generally help an auditor or tester find their way through the tangle of a spreadsheet.

I understand Spreadsheet Auditor no longer exists. But the increasing complexity of spreadsheets and lack of a decent type system in Excel have stimulated development of other tools. Googling for resources, I found Patrick O'Beirne's list at Spreadsheet auditing and inspection tools. Anyway, Ray, you were talking about training Customs staff to use Spreadsheet Auditor.

Yes indeed. I trained over 100 colleagues — not just VAT, because as far as I recall, we had tobacco duty and an oils calculation too — but mostly VAT inspectors, to use it in 1992. This was the only training course in Customs's history to directly collect extra tax, as we used live business data in a series of mass audits — a bit like fiscal Moonie weddings.

That's a nice analogy! It must have made the training courses exciting: a new developer is learning to use the audit program, finds a tax underpayment in a spreadsheet, and can then send round the boys in blue to go and bash down the offending business's door in a dawn raid...?

When I first got Webbed up at work my first AltaVista search was for "spreadsheet" with "error", which found Ray Panko's web site. We exchanged emails — hence Ray citing my programme of audits, written up in my paper Is This Spreadsheet a Tax Evader?, in his research — and he put Dave Chadwick, Pat Cleary and I in touch.

Yes, I remember your study being mentioned in a 1997 New Scientist feature called Fatal Addition. The feature talks about Ray Panko's studies finding a "dangerously high rate of errors", and Dave Chadwick's work with accountancy firm KPMG to help auditors detect errors. And of your audits, it says:

A study by the Computer Audit Unit of HM Customs and Excise also found a relatively low error rate of 11 per cent. But where there were mistakes, the spreadsheets were out by amounts ranging from a few hundred pounds to millions of pounds.

The exercise was so successful in turning up mistakes in the way companies calculate VAT payments due on their sales that all VAT inspectors are now taught to spot flaws in spreadsheets, either in the way they are set up or through wrong entries. Not all spreadsheet errors are simple mistakes. A spokesman for the Computer Audit Unit says that at least one case analysis of a spreadsheet has produced a conviction for fraud.

We met up, compared notes, concerns, preventive and detective control ideas, and kept in touch. I was President of the Information Systems Audit and Control Association Northern England branch, and wearing my ISACA hat, I had suggested that spreadsheet risks would be a good, small scale research area that the branch could fund, initially by sponsoring a symposium. When discussing this with Dave and Pat (in a pub, hence 3 guys in a bar) we thought "Why not do it between us — Let's call an inaugural meeting and see what happens".

We did this in 1999, which is how David Colver, Grenville Croll and other long-time members came on board. ISACA sponsored the first conference with a £500 guarantee against losses (never invoked!). Dave Chad and I both went to HICSS 2000, the Hawaii International Conference on System Sciences. We leafleted mercilessly, and the rest is history. We started thinking we were going to have a symposium — Devil's Dictionary definition: a conference where the speakers outnumber the audience — but ended up with over 40 people at the first event.

I can now claim to be the only Tax Inspector ever to have founded a learned society in order to protect Her Majesty's revenues.

Ray, thanks very much. I'll see you in Greenwich next week. Let's meet for a drink next to the Cutty Sark.