Neil Ferguson’s Imperial model could be the most devastating software mistake of all time

Telegraph ^ | May 16, 2020 | David Richards and Konstantin Boudnik

[grundle]

Imperial College’s modelling... could go down in history as the most devastating software mistake of all time, in terms of economic costs and lives lost.

... those of us with a professional and personal interest in software development have studied the code on which policymakers based their fateful decision to mothball our multi-trillion pound economy and plunge millions of people into poverty and hardship. And we were profoundly disturbed at what we discovered. The model appears to be totally unreliable and you wouldn’t stake your life on it.

Imperial’s model appears to be based on a programming language called Fortran, which was old news 20 years ago and, guess what, was the code used for Mariner 1. This outdated language contains inherent problems with its grammar and the way it assigns values, which can give way to multiple design flaws and numerical inaccuracies. One file alone in the Imperial model contained 15,000 lines of code.

Try unravelling that tangled, buggy mess, which looks more like a bowl of angel hair pasta than a finely tuned piece of programming. Industry best practice would have 500 separate files instead. In our commercial reality, we would fire anyone for developing code like this and any business that relied on it to produce software for sale would likely go bust.

The approach ignores widely accepted computer science principles known as “separation of concerns”, which date back to the early 70s and are essential to the design and architecture of successful software systems. The principles guard against what developers call CACE: Changing Anything Changes Everything.

Without this separation, it is impossible to carry out rigorous testing of individual parts to ensure full working order of the whole.

[NELSON111]

>>>They are only as good as the data you feed into them

This is true - but to say there is no science is stupid. That was my point. We are not talking about random prediction - and you know it. If you don't - then you do not know enough about this conversation to actually be commenting. Go learn some science - get educated - then come back and talk about what you learned.

[9YearLurker]

“Mistake”.

Yeah, sure.

[no-s]

FORTRAN has nothing to do with any problem.

hah, I want to say something naughty but I'm afraid of starting a religious conflict. Also the author is just regurgitating what they heard/read.

However I would like to report that in the 70's and 80's I converted millions of lines of FORTRAN code into APL, mostly models from academics. Typically the APL implementation, including the test suite, was 1/100 to 1/1000 the size of the original. From this I assumed the actual information density of FORTRAN was very low. Of course, the ratio of bugs to LOC is roughly the same across languages, all other things being roughly equal. Often my test suite would discover severe bugs in the original unnoticed by the original authors because they didn't have grant money to eradicate bugs, haha.

[silverleaf]

Too bad about the paywall...

[monkeyshine]

FORTRAN? Isn’t that from the 1970s? Used to program it using punch cards, like ballots with chads in them. Maybe there is an updated version but I haven’t heard that word FORTRAN in years.

[T.B. Yoits]

The model was not a mistake but intentional and this "news" story is an attempt to cover that up.

This is a story of agenda-ists who used the work of snake oil salesman (and there are a lot in the computer modeling world) to push a worldwide economic shutdown to protect their global power.

This fake news does leave one opening though: if enough citizens demand that public policy can cite NO model that doesn't disclose all the raw data, cannot be replicated, and is not confirmed by an entity under the jurisdiction of the public policy owners.

[KarlInOhio]

Criticize the program all you want but the problem almost certainly goes back to three input numbers: "how likely is person A to infect person B?", "during what time period will person B be infectious?" and "how likely is person B to die if infected?" Those three were guesses at the beginning. Doubly so since China and the WHO lied about the disease progression in China.

For computer modelling to work you have to constantly adjust it as you learn real results.

[skimbell]

It was a mistake right enough. Ferguson has been making these same types of mistakes for a very long time. He has missed spectacularly on every projection that he has made.

No one in their right mind would pay his nonsense any attention which is exactly what happened.

A pack of democrats (out of their mind nazis) happened upon Ferguson's CD-19 projection and, desperately needing something to derail the Trump Train, picked it up and started their standard "stompy foot" social engineering.

And here we are.

[monkeyshine]

OK I agree, the language itself isn’t the problem. It’s just odd to see FORTRAN being used these days. I am pretty sure a model could be developed using Microsoft Excel functions that would be just as good. Or any other language or software.

I also tend to agree, the integrity of the modeler is paramount. If you tweak it constantly because the results don’t fit your thesis then that is not scientific. And in the end, when they started this modeling, they really didn’t know all that much about disease spread, susceptibility... hell we didn’t even have many tests. There was no way a model could predict things it didn’t and couldn’t factor.

[Trump.Deplorable]

Yes Sir!

[Pikachu_Dad]

lol.

How many lines of code are in modern software...

Windows 10 is how many lines of (very buggy) code?

[Grimmy]

The boss of a top software firm asks why the Government failed to get a second opinion from a computer scientist

**********

There’s a good part of the problem right there. People still hiding from the reality that the gov is very much involved in producing and perpetuating the frauds.

[Pikachu_Dad]

ah, but did you try Python... (snake it) roflol

[semimojo]

No one in their right mind would pay his nonsense any attention which is exactly what happened.

Who had a better model in January and who should have our leaders been listening to?

[wiley]

I’m more interested in the algorithm than the language that it is written in.

[BobL]

“Imperial’s model appears to be based on a programming language called Fortran, which was old news 20 years ago and, guess what, was the code used for Mariner 1. This outdated language contains inherent problems with its grammar and the way it assigns values, which can give way to multiple design flaws and numerical inaccuracies. One file alone in the Imperial model contained 15,000 lines of code.”

Fortran got our ICBMs to hit within 50 meters of their targets in the Pacific, not to mention the accuracy of Mariner 1, so I wouldn’t, necessarily, denigrate it.

[Cold Heart]

“Ferguson has been making these same types of mistakes for a very long time”

Thank you for that insight.

I have suspected that the original “accidental” release of the Wuhan virus may have a similar origin.

Designed plausible deniability.

[pepsi_junkie]

In the defense industry there was a major battle between the forces of the DOD developed language - Ada - and C++. C++ won by default because Ada isn't supported by anyone so you can't get tools for it. But a lot of DoD code developed in the 90s is still in Ada.

And nobody knows how to work in it anymore. Developers from 25 years ago moved on, moved up, or forget. So people have to try to learn it as they go and figure out what this complex software is doing on their own. Upgrading those systems costs a fortune just for that reason alone.

[DoodleBob]

If FORTRAN was the problem, why didn’t Mariner crash and burn?

[DeplorableGirl]

And I’ve lived thru dozens of hurricanes living in South Louisiana, you people have gotten it wrong, endangered lives with useless hysteria, & caused misery....based on wrong models. It’s a pure guess.
But you never lose your jobs; Damn shame.