The h Bar

12:02 AM

Hey physics people, I was thinking why do we usually use single character variable names in physics, unlike in programming where we use descriptive names? (I know writing m is shorter than mass but at-least it is more descriptive?)

ACuriousMind

12:18 AM

@AwalGarg I think it is mostly for brevity - formulae quickly get cluttered even with one-lettered variables, and longer names would be horrible for writing on boards or by hand in general

alarge

@AwalGarg In physics you rarely have more than a handful of variables, so there's rarely an issue with confusion. Of course in numerical codes, where there are many variables, you'd typically use more descriptive names. I think it is mostly due to mathematical notation and convention. For example this article uses descriptive names in the pseudocode, but single letters in the mathematical analysis.

dmckee

We also use reliable convention for the single variables. In mechanics $m$ will almost always mean mass, and $v$ velocity. The reliability of these conventions gives an opportunity to become used to the brief names.

By contrast, non-trivial programs roam over the concept landscape rather than staying in a single well defined realm. (Of course when you first started programming your programs were trivial by this definition, and you could get away with short variable names).

Awal Garg

@ACuriousMind Well, that's true. But I just found out that if I write my derivations and formulae as functions (like we do in code), it is a lot more readable and easier to understand.

alarge

Now that I mentioned sort algorithms, here's my favorite video on the matter.

Awal Garg

@alarge Ha! I am talking about things like q representing charge (I can't honestly imagine how that letter relates to charge unless some other language thing)...

don't we... like... run out of alphabets?

@dmckee I do agree, but for more complex cases, things kinda get out of hand. For instance, if I have to represent two masses, my teacher recommends us to use m1 and m2... I don't really find it intuitive.

dmckee

12:27 AM

@AwalGarg Thus the greek, and the variaent ways to write the greek, and the caligraphic roman characters and the blackboard bold characters and even a few bit borrowed from Hebrew and Cyrillic.

@AwalGarg Think of it as an array. Indeed when we start working more general case we talk about a set of masses $m_i$ with $I = 1 \dots n$. and then start summing over things.

Of course, thinking is parallel arrays is discouraged by modern programing languages, but hey, I know Fortran 77.

Awal Garg

@dmckee Hmm, I agree. Thanks!

Yeah, basically I had this thought if we could study physics' mathematical side as code. Writing functions, reusing them, and descriptive variable names and stuff. I found the idea intriguing and wanted to know what people here would think about it. So, thoughts?

ACuriousMind

@AwalGarg You can cast any mathematical reasoning into (pseudo-)code. It tends to not be very enlightening.

dmckee

@AwalGarg ::shrug:: Even serious computationalists write traditional formulas when they want to explain ideas instead of implementations. Could just be habit, or there might be something to it. I know that I read math faster in the traditional form than in code.

Awal Garg

@dmckee Fair enough. I guess I just find code easier because of my code-fu.

@ACuriousMind not even readability?

alarge

I find it easier to see "symmetries" in written out mathematical form. For example, I might have a nested double loop in code and not immediately notice that it could be written as a single sum with some clever manipulation, whereas in mathematical form this might be more apparent. I think it's mostly due to my past experience doing lots and lots of algebra, though, and some might find these things easier to see in for loops.

ACuriousMind

12:40 AM

@AwalGarg I don't find code more readable than a mathematical formula, and I suspect most physicists and mathematicians would agree

Awal Garg

@ACuriousMind I suspect the same :P Possibly because of conventions.

@alarge Depends on the programming language. Writing it in ruby would be fairly easy. Probably even better in Haskell.

user54412

1:22 AM

@AwalGarg I think there's something to be said for brevity when it comes to digesting the equation as a whole, and I say this as someone who does at least as much coding as formula writing.

user54412

For instance, grabbing a random equation from my thesis, I have $$\Delta r \geq \sqrt{\Delta_+} - \sqrt{\Delta_-} + M \log\left(\frac{r_++\sqrt{\Delta_+}-M}{r_-+\sqrt{\Delta_-}-M}\right)$$

user54412

This would be a horrendously long line in code with descriptive variable names, or probably it would be broken into multiple lines

user54412

(I should know -- I have to code it)

user54412

And then the symmetry and intuition would not be visible at a glance

user54412

Also, it helps that writing has more characters, more ways to decorate characters (overlines, subscripts, etc.), and more ways to position them (fractions, etc.) -- so information can be encoded in more than just raw letters

bolbteppa

1:28 AM

Any of you guys familiar with deriving the $\dfrac{C}{2}(1 - \cos(\eta))$ solution from the GR Friedmann equations?

Transcript for