Think before you code

I have been reading Flash boys by Michael Lewis. It is a fascinating book about Wall Street and high-frequency traders. I will write a review about the book later but I can't refrain from an overview comment. Apart from the greediness, malice, and sneakiness of Wall Street people, another thing that stood out about them was the pervasive cluelessness and ineptness. It turns out nobody knew what they were doing, including the people that are supposed to be regulating things, even sometimes the people that are gaming the system. This brought to my mind Steve Jobs' quote from his 1994 interview: "Everything in this world... was created by people no smarter than you."

Anyways, today I will talk about something completely different in the book that caught my eye. It is about thinking before coding.

On Page 132 of the book:
Russians had a reputation for being the best programmers on Wall Street, and Serge thought he knew why: They had been forced to learn to program computers without the luxury of endless computer time. Many years later, when he had plenty of computer time, Serge still wrote out new programs on paper before typing them into the machine. "In Russia, time on the computer was measured in minutes," he said. "When you write a program, you are given a tiny time slot to make it work. Consequently we learned to write the code in ways that minimized the amount of debugging. And so you had to think about it a lot before you committed it to paper.... The ready availability of computer time creates this mode of working where you just have an idea and type it and maybe erase it ten times. Good Russian programmers, they tend to have had that one experience at some time in the past--the experience of limited access to computer time."

Of course, Dijkstra was a big proponent of thinking before programming. Here in EWD 1284, he compares European CS vs American CS. He didn't have nice things to say about the keyboard-happy American CS.
The major differences between European and American CS are that American CS is more machine-oriented, less mathematical, more closely linked to application areas, more quantitative and more willing to absorb industrial products in its curriculum. For most of these differences there are perfect historical explanations, many of which reflect the general cultural differences between the two continents, but for CS we have also to take into account the special circumstance that due to the post-war situation, American CS emerged a decade earlier, for instance at the time when design, production, maintenance and reliability of the hardware were still causes for major concern. The names of the early professional societies are in this respect revealing: the “Association for Computing Machinery” and the “British Computer Society”. And so are the names of the scientific discipline and the academic departments: in the US, CS is short for Computer Science, in Europe it is short for Computing Science.
The other circumstance responsible for a transatlantic difference in how CS evolved I consider a true accident of history, viz. that for some reason IBM was very slow in getting interested in Europe as a potential market for its computers, and by the time it targeted Europe, this was no longer virgin territory. Consequently, IBM became in Europe never as predominant as it has been in Northern America. 

Here in EWD1165, Dijkstra shares an anecdote about thinking before coding and uses that as an opportunity to take a shot at "software engineering". Man, Dijkstra had the best rants.
A recent CS graduate got her first job, started in earnest on a Monday morning and was given her first programming assignment. She took pencil and paper and started to analyse the problem, thereby horrifying her manager 1.5 hours later because “she was not programming yet”. She told him she had been taught to think first. Grudgingly the manager gave her thinking permission for two days, warning her that on Wednesday she would have to work at her keyboard “like the others”! I am not making this up. And also the programming manager has found the euphemism with which to lend an air of respectability to what he does: “software engineering”.

In fact, Dijkstra takes things further, and advocates even forgoing the pen and the paper when thinking:
What is the shortest way to travel from Rotterdam to Groningen, in general: from given city to given city. It is the algorithm for the shortest path, which I designed in about twenty minutes. One morning I was shopping in Amsterdam with my young fiancée, and tired, we sat down on the café terrace to drink a cup of coffee and I was just thinking about whether I could do this, and I then designed the algorithm for the shortest path. As I said, it was a twenty-minute invention. In fact, it was published in ’59, three years late. The publication is still readable, it is, in fact, quite nice. One of the reasons that it is so nice was that I designed it without pencil and paper. I learned later that one of the advantages of designing without pencil and paper is that you are almost forced to avoid all avoidable complexities. Eventually that algorithm became, to my great amazement, one of the cornerstones of my fame.
Dijkstra (2001), in an interview with Philip L. Frana. (OH 330; Communications of the ACM 53(8):41–47)"

In several of his EWDs, Dijkstra mentioned how he favored solving problems without pen and paper and just by thinking hard, and how he has the entire article sketched in his mind before he sits to write it down in one go. Here is an example where he mentions the Mozart versus Beethoven approach to composing. Obviously Dijkstra was on the Mozart camp.
There are very different programming styles. I tend to see them as Mozart versus Beethoven. When Mozart started to write, the composition was finished. He wrote the manuscript and it was 'aus einem Guss' (from one cast). In beautiful handwriting, too. Beethoven was a doubter and a struggler who started writing before he finished the composition and then glued corrections onto the page. In one place he did this nine times. When they peeled them, the last version proved identical to the first one.
Dijkstra (2001) Source: Denken als discipline, a program from Dutch public TV broadcaster VPRO from April 10th, 2001 about Dijkstra"

For the record, and not that you care, I am on team Beethoven. Being perfectionist and taking an "I will get it right in one shot" approach to thinking/designing makes me freeze. Separating concerns by dividing my thinking between a creative mode and criticizing mode works much better for me. (I talked about that in here, here, and here.)

Maybe my thinking is sloppy and I need crutches. But it is possible to argue that writing/prototyping is a tool--not a crutch-- for thinking, and that tools are invaluable capability multipliers. And on the power of writing as a tool, I will end with these quotes:
Writing is nature's way of letting you know how sloppy your thinking is. -- Guindon. 
If you think without writing, you only think you're thinking. -- Leslie Lamport 
Without writing, you are reduced to a finite automaton.
With writing you have the extraordinary power of a Turing machine. -- Manuel Blum

MAD questions

1) Hmm, a handicap that becomes an advantage. I think that was the entire theme of the "David and Goliath" book by Gladwell. Of course, that book got a largely negative critical response. But that doesn't mean it cannot be a useful model sometimes.

Are there scientifically proven studies of a seeming handicap becoming an advantage? What are some examples?

2) Yeah, about that whole Mozart versus Beethoven mode thing... Am I mistaken? Should I be more open-minded about the Mozart mode? What is doable by the Mozart mode that would be impossible to do with the Beethoven mode?

3) Well, this is not a question. This is self reflection. Way to go Murat! You started with Flash Boys, jumped to Steve Jobs's 1994 interview, and finished with comparing Mozart and Beethoven mode. A great display of "discipline in thought" indeed.


foo said…
Regarding how much you can do within your mind:

I would say it grows as you have learned more concepts, more methods, more problems and their solutions: the more you learn, the more you can work "in your head". Familiarity, growing reflexes on the object you're studying are key to progress, to the moment of insight where you make progress.

But then, how do you become familiar? By reading, of course, and also by writing (and by explaining, according to Feynman).

Also, how do you keep it all in your head? Some people seem to have larger brains, but also they have greater clarity, often because they see things from the right perspective: in other words, they found a way to reduce the problem to a workable size, often through good notation, good breaking down into smaller problems, good generalization, etc. I'm not sure there is a sure path towards that, but I think that something like "How to solve it" by Polya does help.

So, I think it's a mix of both: at times you do it all in you head, up to a point at which you commit to persistent storage for later processing. In the end, it's a lot about iterating: depending on one's current skills and talents, the size of a step can be smaller or larger, but what matters is that one keeps getting back at it, in a way that converges to a result.
Wander said…
I find myself more on the Mozart side, though perhaps that's no surprise since I'm a countryman of Dijkstra. Like 'foo' said before, I think that an efficient way to keep stuff in your head helps a lot, but it also helps to be able to fetch from "persisted memory".

Also, there are actually three stages in my workflow:
- Thinking inside my head.
- Putting it on paper/github issues/an empty text file in pseudocode.
- Actually "programming" it in a way that a computer would understand.

Popular posts from this blog

Foundational distributed systems papers

Your attitude determines your success

Progress beats perfect

Cores that don't count

Silent data corruptions at scale

Learning about distributed systems: where to start?

My Distributed Systems Seminar's reading list for Fall 2020

Read papers, Not too much, Mostly foundational ones

Sundial: Fault-tolerant Clock Synchronization for Datacenters

I have seen things