- The Art Of Doing Science And Engineering
- Richard Hamming
- personal rating
The goal of this book is to teach the art of thinking about science and engineering. Hamming argues that this art is not taught in the school curriculum, but is crucial to being a successful practitioner.
- Hamming recommends doing quick modelling to verify the truth of various claims, for example using plausible models of the human population to relate the plausibility of 2 independent but correlated claims together.
- One should take their own opinions and try to first express them clearly, and then examine them with counter arguments, back and forth, until they are clear as to what they believe and why they believe it.
- The book was slightly painful to read on the Kindle. And a lot of the content is covered in MacKay’s fantastic information theory book, which makes the mathematics less tedious.
- Hamming distinguishes creativity and originality: there is a
implicit desire for value in creativity. How does one boost
creativity? Hamming doesn’t know, but he identifies several traits:
- false starts and solutions sharpen the next try
- temporary abandonment is a common phase
- “if I had a solution, what would it look like?” is a good question to ask
- Hamming’s advice for “experts”:
- Don’t automatically reject every crazy idea, but also don’t pursue all of them
- Be conscious of your choices, and be wary of experts
- Hamming on unreliable data:
- Hamming’s rule: 90% of the time the next independent measurement will fall outside the previous 90% confidence limits
- You get what you measure: the way you choose to measure things controls to a large extent what happens. This may result in bias.
- Hamming on research:
- He oft repeats this quote from Pasteur: “Luck favors the prepared mind”.
- It is important to believe you can go on to do great things.
- Have drive, and desire excellence. Intellectual interest is a compound interest.
- Take time on a regular basis to ask the larger questions.
- Great people can tolerate ambiguity, both believe and disbelieve at the same time.
- 3 essential skills (selling the idea through clear presentation):
- Giving formal presentations
- Producing written reports
- Giving informal presentations as they occur
- Errors in continuous signals are amplified, but in discrete signalling using repeaters rather than amplifiers, noise introduced is removed if the repeater is able to correct for it. Do we have an internal error-correction mechanism when sending discrete binary signals across neurons?