VisiCalc: The 30th Anniversary | Hacker News Reader

VisiCalc: The 30th Anniversary

pcmag.com

8 pointsastrec8 comments

What a dope Dvorak is.

His theory that the mortgage crisis was caused by spreadsheets making bad decisions is just a little stupid. Management has the company spreadsheets designed to make decisions based on the rules they set down for giving a loan. Those rules are the same whether it's a spreadsheet filling in the numbers or whether people are calculating the numbers manually.

Mr. Dvorak doesn't seem to grasp the concept of automation. Automation simply takes a process and lets the computer do the parts it's capable of. It does not invent a new process. So the loans being given would be the same no matter what.

I think his point is that spreadsheets allow people to produce convincing models (even convincing themselves) without necessarily understanding them. It's like a chainsaw could trim your hedge in no time, provided you managed not to saw through your own legs in the process.

Certainly if Joe Spreadsheet User doesn't understand concepts like numerical stability, he really shouldn't be allowed to make decisions with other people's money.

But again that has nothing to do with the spreadsheet it has to do with the process (or model in your terms). So to give an example lets say the spreadsheet was never invented. That wouldn't have stopped people from creating a mathematical equation to determine credit worthiness. Bankers would just have Joe Spreadsheet do the equation by hand.

He's still going to be equally uninformed and he's certainly not going to be able to verify the validity of the equations while he's doing it. He's simply completing the required steps to get the equation to produce an answer.

So again the spreadsheet isn't obscuring anything that wouldn't have been obscured without it. If anything it's saving us from the stupid mistakes that Joe Spreadsheet would be making if he tried to do by hand what a computer does.

But the algorithm, the equation that Joe is using will have been created by a mathematician or economist in that case. The thing with spreadsheets is you can tweak them 'til it "looks right" with your sample data, but still have no idea how it will behave when exposed to real data. A model built empirically by one untrained person will show them what they expect to see, or worse what they want to see. Maths should be left to the mathematicians just as chainsaws should be left to the lumberjacks.

The algorithm would have been created by a mathematician no matter what. Spreadsheets don't create equations to judge credit worthiness they only execute them. Because computers don't create process they just automate it.

So the difference between the equations being used to calculate credit worthiness should be unchanged whether you're using a computer or a pencil and paper. The computer just does it faster. That in turn means the numbers can be tweaked no matter what.

If anything, the computer makes it harder to tweak the number because a manager who oversees you only has to check the 6 numbers you put in and not spend an hour redoing your calculations by hand.

The reality is mathematicians wouldn't be giving loans no matter what. Just as physicists don't run power plants they just make a procedure and pass it off to a lower paid laborer. So Joe Spreadsheet would be the one giving the loan regardless.

Sadly, this is just another Dvorak troll with no useful information. I would have loved (seriously, that is why I clicked) to read a detailed overview of what VisiCalc did, what it couldn't do, screenshots, features, limitations, quirks, effects on the industry, etc. Instead we get a "man, spreadsheets suck" essay, if one can apply that word.

"This is what caused the mortgage crisis: Spreadsheets instead of people were making decisions on loans."

That's an interesting perspective. It may be that a loan officer's nonquantitative intuition based on personal experience was a better basis for making loan decisions than running a formula on a spreadsheet.

But the real question, which Dvorak misses, is whether management types would trust that type of intuition. Loans are big business so I don't see a company like BofA trusting the intution of individual loan agents. Sad but true.

So I have to believe the process of giving a loan would be based on the same rules even if there were no spreadsheets.

Reply on news.ycombinator.com