Interesting article about how scientists are beginning to abandon algorithms based on natural phenomena:

http://gizmodo.com/scientists-are-turning-their-backs-on-algorithms-inspir-1708656813

 

I disagree with the article.

Einstein and Bohr abandoned natural law 100 tears ago.

I just realized I typed tears instead of years.

After thinking it over, tears is more appropriate.

I have often said that an automotive engineer should not be allowed to touch a pencil, until 5 years of auto repair.

Physicists should not be taught math until after graduation.

Math can not skirt law. Try all you want.

Nature accepts no substitutes.

 

My reading is that they are not abandoning the algorithms, only the terminology that describes them as biological phenomena instead of math. I can see why and where that might help, but I can also see where, if you're working on a problem that resembles something in nature, it's darn handy to find the relevant math by its "common name" instead of its pure math.

For instance in marketing, the diffusion of innovation that often defines a product life cycle is very similar to an epidemic. The "epidemic formula" can fit real marketing data amazingly well. That connection was only made in the first place based on a hunch that the epidemic model might be meaningful. It was. I'm not sure a marketing guy would have ever made that connection without the biological clue.

 

Interesting article about how scientists are beginning to abandon algorithms based on natural phenomena:

http://gizmodo.com/scientists-are-turning-their-backs-on-algorithms-inspir-1708656813

Using "algorithms" is far too broad a term to use in this context. It is a particular type of algorithm, namely genetic algorithms (what they called evolutionary algorithms), that is being discussed. Even then, they are talking about changing the terminology used, which seems rather pointless to me.

The point I latched onto was near the end where they said, "We are also looking at introducing rules from software engineering known as formal design patterns, which essentially set down prescribed ways of solving a given problem to stop people constantly trying to come up with radical alternatives."

That seems completely counter-productive. The purpose of those rules in software engineering is to constrain the solution space to well-known and well-understood methods. Why on Earth would you want to constrain the solution space of a problem that has no known solutions at all!

 

Why on Earth would you want to constrain the solution space of a problem that has no known solutions at all!

And why on earth would anyone want to put a limit on creativity?

 

And why on earth would anyone want to put a limit on creativity?

There are lots of times that you want to do that -- the software engineering example is but just one.

Imagine taking your car in to get your brakes serviced. Do you really want to go to someone that is going to be "creative" in how they do brakes, or do you want to go to someone that uses industry best-practices in how they do brakes?

Creative in R&D is great, creative in production is generally not. This is not to say that creativity is never applied to production methods, but when it is it is known as "production R&D".

And that's really the point I was trying to make here.

Developing algorithms to solve these problems is R&D and you want to encourage unfettered creativity.

Software engineering is about using best-practices to write production code.

 

There are lots of times that you want to do that -- the software engineering example is but just one.

Imagine taking your car in to get your brakes serviced. Do you really want to go to someone that is going to be "creative" in how they do brakes, or do you want to go to someone that uses industry best-practices in how they do brakes?

Creative in R&D is great, creative in production is generally not. This is not to say that creativity is never applied to production methods, but when it is it is known as "production R&D".

And that's really the point I was trying to make here.

Developing algorithms to solve these problems is R&D and you want to encourage unfettered creativity.

Software engineering is about using best-practices to write production code.

You're right... I should've been more specific... Of course I was referring to R&D and when trying to solve situations that have not been encountered before... such as when NASA's engineers devised that clever plan to bring the Apollo 13 astronauts home. That was pretty creative of them.

I agree that most everyday known situations are definitely best dealt using a by-the-book approach to fixing them... such as your car's breaks and steering system.

 

Apollo 13's a good example -- and I wouldn't have thought of that as R&D but, in this context, I guess that's a pretty fair description.

NASA, at least at the time, spent a significant fraction of their time throwing crazy problems at everyone specifically to accomplish both objects -- force them to know the procedures inside and out and to develop their ability to think outside the procedures and do so quickly. It paid off.

 

Apollo 13's a good example -- and I wouldn't have thought of that as R&D but, in this context, I guess that's a pretty fair description.

NASA, at least at the time, spent a significant fraction of their time throwing crazy problems at everyone specifically to accomplish both objects -- force them to know the procedures inside and out and to develop their ability to think outside the procedures and do so quickly. It paid off.

I actually read the book... and although the movie wasn't that bad (except for the idiotic part in which Mr TM fires up the LM and drives it manually), the book (as is often the case) was waaaaaayyyyy better!

 

I thought the book (I'm assuming you are talking about Lost Moon) was a horribly written book. That it was a compelling read was due to the subject matter and not the writing. It also couldn't make the proper distinction between voltage and current.

Who is Mr TM? Are you talking about the PC+2 burn, or after the initial power-up of the LM after the explosion?

 

I thought the book (I'm assuming you are talking about Lost Moon) was a horribly written book. That it was a compelling read was due to the subject matter and not the writing. It also couldn't make the proper distinction between voltage and current.

Who is Mr TM? Are you talking about the PC+2 burn, or after the initial power-up of the LM after the explosion?

MY BAD... I meant to write TH... Tom Hanks!

And yes, the reason I liked the book so much was not because of its narrative, but because of its excellent technical detail.
But, if you want to have it both ways regarding space exploration... I strongly suggest you try this book.

 

I'll add that to my reading list -- assuming I every actually get back to it.

 

Back in the day, I worked for a ex NASA project engineer. He told me of a study that NASA conducted on creativity verses employ-ability. The charted data showed that as the persons creativity increased so did their employ-ability, but only up to a point. When a person got too creative, they lost the ability to be at work on time, go to meetings, meet schedules, etc. and their employ-ability dropped like a rock. I think this was a short sighted study. If a person is too creative, they just need a good secretary.

 

As with most things, there needs to be a balance. I've known people that were exceptionally creative, but their ability to stay within bounds of reality were virtually non-existent. On the other hand, I've known people that were extremely grounded in reality, but their utter lack of creativity made them worth little more than a good automaton.

 

There's an article in the news I just heard about showing a connection between creativity and psychotic tendencies. I agree there's a sweet spot, certainly for people that need to work in a corporate environment or with a team.

Actually, I just saw "The Imitation Game" recently and in part it dealt with Alan Turing's difficulties in working on a team.