Comments on: Feynman’s Rigor http://jsomers.net/blog/feynmans-rigor the jsomers.net blog Sun, 29 Jan 2012 20:57:13 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: James Somers http://jsomers.net/blog/feynmans-rigor/comment-page-1#comment-5 James Somers Tue, 03 Mar 2009 04:32:51 +0000 http://jsomers.net/blog/feynmans-rigor#comment-5 <p>Good points. I think my mistake here was in having <a href="http://projecteuler.net" rel="nofollow">Project Euler</a> in mind while writing. As you know, PE is a special kind of introductory programming in that the problems there require a single, precise answer. So the inflexibility that makes programming in general a (relatively) good exercise in rigor is really amplified.</p> <p>When I consider e.g. web programming, I agree with you that you can get away with a lot without deep understanding; but I still find myself tracing all kinds of errors in detail. As someone on the Internet said (in a comment somewhere, I think):</p> <blockquote>[on programming]: I don't know of any other discipline, even math, where one is more repeatedly confronted with one's mistakes.</blockquote> <p>That seems roughly true. And the thing about mistakes is that they <em>force you</em> into the kind of thinking that I imagine Feynman doing, which is buried in the details.</p> Good points. I think my mistake here was in having Project Euler in mind while writing. As you know, PE is a special kind of introductory programming in that the problems there require a single, precise answer. So the inflexibility that makes programming in general a (relatively) good exercise in rigor is really amplified.

When I consider e.g. web programming, I agree with you that you can get away with a lot without deep understanding; but I still find myself tracing all kinds of errors in detail. As someone on the Internet said (in a comment somewhere, I think):

[on programming]: I don’t know of any other discipline, even math, where one is more repeatedly confronted with one’s mistakes.

That seems roughly true. And the thing about mistakes is that they force you into the kind of thinking that I imagine Feynman doing, which is buried in the details.

]]> By: Avinash http://jsomers.net/blog/feynmans-rigor/comment-page-1#comment-3 Avinash Tue, 03 Mar 2009 03:58:09 +0000 http://jsomers.net/blog/feynmans-rigor#comment-3 <p>While I think you were just trying to make a admittedly correct point, an interpreter is called as such because it interprets from one form of code to another--and yes, it interprets. The reason it does so exactly is because it is doing so correctly.</p> <p>I would argue very, very strongly that you <strong>absolutely</strong> can produce something that you don't understand while hacking. Scaffolding is almost by definition exactly that. I can make a fairly complex web application in Django, Ruby on Rails or CodeIgniter, but there's no way I know the details that are going on.</p> <p>I would even argue that a lot of code is produced out of dumb luck--I know that has happened to me on several occasions, and I've read more than enough production-level code to know that it's not uncommon.</p> While I think you were just trying to make a admittedly correct point, an interpreter is called as such because it interprets from one form of code to another–and yes, it interprets. The reason it does so exactly is because it is doing so correctly.

I would argue very, very strongly that you absolutely can produce something that you don’t understand while hacking. Scaffolding is almost by definition exactly that. I can make a fairly complex web application in Django, Ruby on Rails or CodeIgniter, but there’s no way I know the details that are going on.

I would even argue that a lot of code is produced out of dumb luck–I know that has happened to me on several occasions, and I’ve read more than enough production-level code to know that it’s not uncommon.

]]>