archives

Let's fix spreadsheets

I wonder: why is it that spreadsheets appear to be ignored by real programmers?

Is it because spreadsheets are the equivalent of spaghetti code? Is it because spreadsheets are hacks, put together by non-savvy 'business' people?

The horror!

Yes, I admit. I also very much believe all this to be true. Spreadsheets are sloppy programs. Hence, we programmers don't want to bothered by them. At the same time, research shows that almost 5% of the spreadsheets contain serious errors, causing multi-million dollars of losses.

That figure should freak you out, if you happen to own a bank account. I know firsthand that our banks still run some of their critical businesses on broken spreadsheets. The same spreadsheets we programmers tend to ignore.

Sure, there are attempts to lift spreadsheets into the realm of functional programming, with Reactive Functional Programming (FRP) as the most recent development. But FRP doesn't feel right. As a programmer? May be yes. But I don't think the 'poor' people will understand. FRP is just to distant from their beloved spreadsheets.

Can we educate people to learn FRP? I don't believe we should. What I believe is that we must teach them appropriate 'programmer values':

Modularity
Expressivity
Repeatability
Testability
Control
Performance
Versioning
Continuous builds
Production runs
No crashes please

None of these can currently be attributed to legacy spreadsheets. Ultimately, that's why I think legacy spreadsheets are bound to be a failure. We need more programming, without getting in the way of 'getting things done'.

So that's our challenge: to enhance spreadsheets, so that programmers and business people will meet each other in cells. It is obvious that I've accepted this challenge: to develop a new powerful spreadsheet paradigm. A paradigm for everyone to like.

Edit: a small edit