On comments
Conversation activity · last 35 hours peak 3/30m
Clustered from 18 items across 2 sources. Not yet parsed — the coverage below is the raw record.
Social posts 2
Voices from the web unedited
-
There's no good answer, at least not one anyone's willing to commit to. There are a few projects that try to track issues directly with source control, inheriting a few extra features from Git, like having them all work offline, having every laptop be a backup of the entire project, and being able to easily port them all into whatever you're going…
-
I believe, that the book works in its entirety, not taken apart quote by quote. Well, since the book contradicts its own dogmatic prescriptions repeatedly, the only way it can work is if the reader treats it like a puzzle to be assembled into some coherent meaning. But that’s not what most people would consider the hallmark of a clear and…
-
You should not grimace and you should not feel the failure unless you're into that. How many devs you've seen that justified their sloppiness in any way possible? If you could be writing a better code; do it. You shouldn't be happy about writing a shit code; therefore you shouldn't be happy about writing comments in lieu of writing a good code…
-
i think one thing that's often overlooked is that comments can serve as a kind of institutional memory for a codebase, helping to preserve knowledge and context that might otherwise be lost over time, like why a particular design decision was made or what assumptions were underlying a certain implementation choice. this can be especially important…
-
By this logic, I'd rather not have documentation. LLM generated documentation is incredibly verbose and over-complicated, instead of just providing 4-5 template files and telling the user to symlink the one they want, it wrote 200 lines of code and 400 lines of docs to automatically generate the resulting file from a single template. The docs are…
-
The author also said I am writing for an audience who I assume is "skilled in the art" I may reply to him later, but I don't feel like it at the moment. I really don't want comments in a function to 'teach' me something. When I'm looking at code I want to confirm the implementation is correct, not what the author thinks is correct
-
Great writeup, especially with documentation generation tools like JSDoc or Swagger, it's nice to tie the "staleness" of comments and doc together so there's some incentive to keep it maintained. Bummed to see that this is on the blog of some AI defence contractor, though...
-
/* * increment the variable in this scope which has the * identifier 'i' by one, using the post-increment * postfix operator * see: https://en.cppreference.com/cpp/language/operator_incdec */ i--;
-
Comments in code are deemed "mostly useless" these days. Says who? I know that this is just a hook for your post arguing about them being useful, but this still irks me 🙃
-
I will als helpfully add "don't increment j here", because it misunderstood the code and wrongly incremented j instead of i the first time.