How to learn programming - Advice for scientists | Clara Sousa-Silva and Lex Fridman

Please support this podcast by checking out our sponsors:

GUEST BIO:
Clara Sousa-Silva is a quantum astrochemist at Harvard.

PODCAST INFO:

SOCIAL:

[stox]

About the productivity trap:

I don't even spend 90% of my time improving my code, I spend 90% of time improving my Linux system like an Oldtimer car enthusiast and 9% of my time I spend on stack overflow to tell other people how bad their code is without providing actual improvements, like every regular user on this forum. With the last 1% I actually get some work done and earn my money at burger king.

[josephgaviota]

As a fast typist, and as a programmer for 40 years, I agree that well written, well commented code should be fairly easy to revisit in the future, without documentation.
My motto: Never sacrifice keystrokes for clarity.

[biggop]

Lex hit it on the nail. Ive made the mistake of focusing too much time on optimization.

[matthewg]

Sousa-Silva - "I would much, much faster hire someone who knows programming but barely knows where space is than teach programming to an astronomer."
That is incredible. Really speaks to the difficulty of writing code.

[MrChaluliss]

I frequently question the amount of meta note taking I do while coding and while learning new concepts related to complicated project tasks. Comments like these from Clara really make me feel like I am not wasting my time though. When I want to go back to anything I have done, it usually doesn't take long to catch onto the whole picture of it so far, as I often am explicitly documenting the purpose of each object, action and task along side actually doing the work. Human memory is tricky

[Nick-kbjc]

Bad programmers are usually the people who decided to do it for the wrong reasons. I know many people I went to university with who were only CS majors because they heard you could make a decent salary with a CS degree. They were never invested in learning data structures and algorithms the right way and didn’t like coding and making stuff in their free time. To truly be a good programmer, you have to know data structures and algorithms, and also like to code when you don’t have to. For me, it was making video games that kept me motivated.

[wiskasIO]

I guess that depends on the person, I have a friend that is a Nuclear Scientist and he only knows one language (FORTRAN) but he knows it by heart because he loves speed. He refuses to use other environment or language, he told me that his compiler and the way he codes his Fortran is so optimized that anything that wasn't that fast with his formulas would drive him nuts.

[timtags]

I still know as little about code as when I started the video but I did learn not to be obsessed over perfection.

[jsr]

Her analogy that she would rather teach a programmer where space is, is spot on. Because, well space is there. There aren't 10, 000 different form of space above us. But there are 10, 000 different ways to code a single thing.

[caleb-hess]

Great programmers hold "getting the job done" much higher than having to use the perfect methods and tools.

[SuperYova]

There are only 3 things [to programming]:

1. Reading from a variable
2. Writing to a variable
3. Conditional branching

Everything else is syntactic sugar.

[ridebywire]

You wont learn programming in shool. You learn it at home. Find something that interests you. Music? Develop a music player. Games? Develop a small game with sprites. Thats the start. Then apply what you learn at shool.

At shool, we should develop a program to calculated taxes. Where is the fun in that? Instead I have started with C++ and played around with a small game engine. Thats how it started for me.

[GodofStories]

I love Claras confidence and strong opinions, a brilliant person!❤

[garystroup]

I think for any programmer in general, taking the time to write comments about what the block of code is doing will save some of the frustration for you and others who have to read your code in the future.

[dariuszgadugadu]

Read Clean Code. Strongly recommend it.

[GandalfTheBrown]

Wow she is really well-spoken. Like her live sentences sound like writing

[aftalavera]

Lex you are amazingly wise for your age!

[marcvanleeuwen]

Curious to say, just after hearing that people are still programming in Fortran because of legacy code, that in reality code (whether well written or not) will most likely very quickly become obsolete.

[tommyhuffman]

You choose such fascinating topics to cover! Good talk.

[dannyiskandar]

get the job done first, i guess is more important later when there is time to revisit the code then improve it. I think that's my approach in life. In general it is better to be early than late? Probably.