Erik Meijer – Patents, products and papers

I am visiting the European Computer Science Summit today, to give a presentation on social media for scientists (more on that later) First, there is a keynote by the infamous Erik Meijer! After a short morning snack of category theory:

Erik starts his real topic: how to transfer knowledge out of universities. Currently, the method of choice is patents. Erik Meijer says that people view patents like magic exercise machines: just writing a patents makes you rich, just as using the machine for 10 minutes a day makes you skinny. This is not true. Patents are expensive and not so effective.

First of all, it can take years for them to be approved, and that is of course added to the time it takes to invent them and sell them. This is not so easy, says Erik. We aren’t even that good at selling our ideas, so selling patents are even harder.

Then, it takes money to apply for them and to enforce them (i.e. start a law suit)

Furthermore, one little patent is not going to help you. Big companies have huge portfolios of patents and most likely there is one that is similar to yours.

Finally, it is very hard to detect a violation. Consider for instance a code generator. How to detect when code is generated with a generator similar to yours.

So what then?

If you want to have high-leverage investment, why not become a venture capitalist? Erik argues that this would be a  better way to spend your money. First of all ‘Silicon X’ is not going to work. You cannot recreate Silicon Valley in your country or region. Why? There is a huge concentration of IQ and risk is low (as you can go to the next start-up, which will be right across the street) Dutch initiatives like YesDelft do not work, as they are located outside of the city.

Open source to the rescue

But it is not all sad news, as there is open source. Be generous up front, and later you can reap the rewards.The idea is that everyone works together on the infrastructure, and then everyone can benefit from services on top.


This poster can be found at Facebook.

Erik argues that research : startups as regular chess : quick chess. In research you have to be slow, think twice, whereas in startups you have to think quick, make mistakes and recover from them.

Erik added more comparisons:

20131009_103443The whole system of publishing is bad, as many many papers are boring, little delta’s. Erik argues we should do more crazy things. Researchers are missing the boat!

GIT, noSQL, Big Data, Machine Learning etc.

This should have come from us! Now, after the fact, researchers are trying to formalize this, but practitioners don’t even read this. Everybody knew that relational databases were crap, but research did not do anything about this.

At POPL, you can hardly sell a paper on dynamic typing, but the whole world runs on JavaScript.

Physicists are currently working on quantum computers, but we are not there with them! They are now designing programming languages, and Erik predicts they will build quantum Fortran and then computer scientists will be complaining.

Thought provoking stuff!


1 Comment

  1. Neil

    I think the most common form of knowledge transfer is via students/researchers getting hired at a company.

Comments are closed.