Modeling How Programmers Read Code | Hacker News Reader

Modeling How Programmers Read Code

synesthesiam.com

73 pointsjoeyespo34 comments

This highlights how "good" programmers are simply those who are most able to think like a computer. The article's language further reinforces this (perhaps intentionally), e.g. the reference to "compiling" functions.

I find this contraint interesting - I don't know if computational thinking is a strength or a weakness. Increasingly, it seems that computational models occur naturally and therefore the ability to think in such a manner would have inter-disciplinary value.

If we deem it a weakness, then programming becomes a UX problem rather than language one. The lack of change both within and across programming paradigms would suggest that many don't believe this to be a fundamental issue.

This actually touches on Alan Kay's statements about programming with "what" instead of "how". It has made me seriously consider logic programming for the applications I write.

I'd place my skill bars higher: "thinking like a computer" seems fundamental: I don't think one can even be a mediocre programmer without it, and it's not enough to make you a good programmer.

A good programmer also needs to be able to effectively communicate through code to both others and themselves, know how to design maintainable programs, how to preemptively avoid bugs by making their code harder to misuse, and too many other skills to list here -- and not all of those naturally flow from knowing how to think like a computer.

Computers don't think like computers. Computers reason formally. Thinking like a computer is just a matter of learning to reason formally and exactly instead of employing fuzzy mental abstractions.

This is primarily hard because we're wired for fuzzy mental abstractions, since they make loads of tasks (natural language, making breakfast, manual labor) far easier compared to taking genuine formal approaches, and we have a lower layer of the brain that can be trained to follow exact procedures quite well anyway.

With more research along these lines, we may have a day when a programming job interview will consist on interacting with some codebase for 10 minutes, and the model will spit out our normalized scores on syntax comprehension, algorithmic thinking, library familiarity, etc.

It does sound representative of the bulk of the daily work.

has there been any work on this? it seems relatively straightforward to score a git branch in certain time scales

EDIT should have RTFA before commenting, gaze tracking stuff is very cool

I wonder if when we get that far the computer is able to write the code for us anyways and programmers will be redundancies. Its just a matter of if, not when. Understanding human cognition well enough to evaluate it is a good indicator that the singularity is near.

Computers already write code for us, they just need to be told what we want them to write. The question is if we can design a programming environment where the computer can understand 'natural' human specifications and compile that, instead of needing to have a trained human compile the natural specifications into source code, which the computer can compile into a program.

Sure we could. This is just like Siri, you tell it what you want, she asks for clarifications, a conversation you have with the computer until...bam, you get what you want.

I would think we could do much of this already if we tried, but there are a lot of things on my list to do until I get to that.

Your optimism on this seems unfounded in my opinion.

When there are no domain restrictions, computers have not proven very capable of comprehension. Look even at the comparatively simple domain of handwriting recognition.

More importantly, you entirely miss the problem that many times the human doesn't know exactly what he wants until he sees it.

As a software engineer, I have no fear my job will be replaced by computers talking to product managers.

That's why the program should be expressed as a conversation! We don't know what we want so we should start vague, get feedback, provide clarifications, and so on. The conversation with the computer must be a two-way thing!

That's brilliant. Not in terms of some AI working out what a silver haired old grandmother meant but in terms of talking to clients - they could imagine talking to a computer and I am just providing that feedback.

Daily or weekly, feedback nice and quick

The thing is, we probably don't need hard AI yet to do some of this. Yes, it must be a dialogue system, but we have those today. We already saw the movement in this direction with Hugo Liu's concept.net work, but for some reason no one has followed up yet. We are getting to the point with speech rec/understanding technology that someone is bound to try again soon.

In general, we don't need Hard AI for much anything, and usually don't want it. Any specific problem you want solved will require some specific solution method or algorithm rather than the whole and entire "make a computer hold an adequate water-cooler conversation" thing.

I think it's less worthwhile to argue this than to try to build it.

That's kind of how TeX and METAFONT have worked since the 1970s. But maybe you have something else in mind.

The problem is that most people are simply incapable of giving correct specifications.

In a lecture about this topic (how to create specifications that can be turned into formal correct code) we were given the following simple example of an incorrect specification:

"Everybody loves my baby. But baby loves nobody, but me."

If you formalize this, you can simply conclude that the person that says that is equal to the baby - which clearly is not, what you intended.

And this is a very, very simple specification. Specifications of real software are magnitudes more complicated.

That is why the computer must hold a two way conversation with the programmer. How often do we just type out a program from our head to the computer anyways? Most of us write the program, debug it, change it because the result wasn't right, or we didn't really understand what we wanted, or someone else decided the requirements changed, or whatever...

A large chunk of the job of developing software is getting the stakeholders to understand the problem they want the software to solve. The rest of it is just typing which is the trivial part.

Yeah, right. And all the technical books on design patterns, functional programming, algorithms, etc. are out there just to teach programmers how to type faster ...

That was the hype surrounding Fortran when it came out in 1959: it would eliminate the need for programmers, because you could just punch the formulas and it would translate them into a program automatically. And in a sense it did, and its successors did to an even greater extent. Today you have "nonprogrammers" who can build, in a day, in OpenOffice Calc, computations that would have taken weeks to get written by programmers in the 1950s (when they could have been done at all.)

It's partly a matter of giving the objectives to the computer at progressively higher levels, and partly a matter of improving user interfaces so that you can tell what the thing is doing and what it's going to do.

These predictions always fail to account for C++.

That is, having computers write code depends on the code being in a form about which computers can formally reason. If you can't prove anything about your code because the language is that bad, good luck getting a computer to write code.

/explain_the_joke.jpg

I would love to see this kind of study done comparing different programming paradigms (i.e. functional vs. procedural vs. OO vs. logic). Some concrete cognitive data on how lisp-like languages may or may not actually make programmers more productive would be quite fascinating I think.

I’m not aware of many eye-tracking studies specifically, but there has been a lot of research done on the general subject of how we read and understand code. Search for papers on “program comprehension”.

To address (some of) your specific point, I suspect part of the appeal of functional programming languages is that they naturally promote describing the data flow of the code. In contrast, imperative languages inherently dwell on control flow, even though it’s often an unimportant implementation detail. It turns out that even when we’re reading imperative code, we’re trying to figure out the underlying data flow anyway, so why not cut out the middle step?

I also suspect part of what holds back functional programming languages from wider acceptance is that when you’re modelling something where time/order matters, the control flow is not an unimportant implementation detail. You have to work harder to describe it in a functional language where you get a lot for free with an imperative programming language.

> You have to work harder to describe it in a functional language

This isn't necessarily true. In Haskell you can use do blocks like:

main = do

action1

action2

...

and your actions will happen sequentially. This is just syntactic sugar behind "action1 >> action2". The real reason for the special syntax is when the actions also return a value you can do "x <- action1", and use the variable x. This is still just sugar, but the unsugared version gets messy pretty quickly.

The big difficult in going from imperative to functional languages is thinking about data structures as immutable. You go from asking how do I change the value to asking giving this value, how do I construct a new value.

In Haskell you can use do blocks like:

Monads in Haskell were exactly the example I had in mind when I wrote my previous post. As you say, you can use ‘do’ notation in simple cases, but it’s just syntactic sugar and the unsugared version quickly becomes unwieldy if you need more flexibility.

Edit: A good case study is the not-quicksort quicksort that gets used all over the place as a demonstration of Haskell’s elegance. Something closer to a real quicksort is rather less tidy: http://www.haskell.org/haskellwiki/Introduction/Direct_Trans...

The big difficult in going from imperative to functional languages is thinking about data structures as immutable.

That’s certainly one of the big jumps, but I don’t think it’s the whole story. You can do a lot with purely functional data structures, given a bit of time to get your head around them, but sometimes you might still have good reasons to work with mutable data instead. That can be very clumsy when shoehorned into a primarily functional programming language.

I’m hoping that one of the next big steps forward in programming language design will be providing much better control of externally observable behaviour, but without forcing a purely functional style or requiring prohibitive amounts of boilerplate code. I’d like to program in a style that is declarative by default and provides powerful tools for working with functions, but also with the ability to break out of that immutable world in controlled ways as easily as writing imperative code and as safely as writing code with a good type system today.

> As you say, you can use ‘do’ notation in simple cases, but it’s just syntactic sugar and the unsugared version quickly becomes unwieldy if you need more flexibility.

I've actually found that it's much better to compose monadic functions with (<=<) than use long do blocks. It forces you to structure your code much more appropriately.

My "Brain-Clearing Whiteboard" has a scrawl on it about investigating "refactorability" as an empirical code metric. I'm taking this Hacker News convo as raw material for figuring out what that would constitute, but broadly speaking my hypothesis is that "more refactorable is better".

I see how this is interesting but I can't get over how the author says, "When contrasting this with Eric's video", but "Eric's video" is a different program than the novice programmer. Maybe this is me being too pedantic.

Not just a different program, a better one, and easier to understand.

It especially annoyed me when he praised Eric for going back to understand the function first, when that function isn't even present in the novice's program.

Previous related discussion:

https://news.ycombinator.com/item?id=4940952

In light of recent mass surveillance uproars, the possibility of remote eye tracking technology is what really scares me. While the technological usage could be limited or masked, such attention detector and analyzer could give enourmous powers to those who have the behavioral models.

Very interesting reserach though.

> he "compiled" the between function in his head

Really? How about "adequately understood the function and did not have to reread it". We are humans, not compilers. ;)

Reply on news.ycombinator.com