[D] Why can't you guys comment your fucking code?

Seriously.

I spent the last few years doing web app development. Dug into DL a couple months ago. Supposedly, compared to the post-post-post-docs doing AI stuff, JavaScript developers should be inbred peasants. But every project these peasants release, even a fucking library that colorizes CLI output, has a catchy name, extensive docs, shitloads of comments, fuckton of tests, semantic versioning, changelog, and, oh my god, better variable names than ctx_h or lang_hs or fuck_you_for_trying_to_understand.

The concepts and ideas behind DL, GANs, LSTMs, CNNs, whatever – it's clear, it's simple, it's intuitive. The slog is to go through the jargon (that keeps changing beneath your feet - what's the point of using fancy words if you can't keep them consistent?), the unnecessary equations, trying to squeeze meaning from bullshit language used in papers, figuring out the super important steps, preprocessing, hyperparameters optimization that the authors, oops, failed to mention.

Sorry for singling out, but look at this - what the fuck? If a developer anywhere else at Facebook would get this code for a review they would throw up.

• Do you intentionally try to obfuscate your papers? Is pseudo-code a fucking premium? Can you at least try to give some intuition before showering the reader with equations?

• How the fuck do you dare to release a paper without source code?

• Why the fuck do you never ever add comments to you code?

• When naming things, are you charged by the character? Do you get a bonus for acronyms?

• Do you realize that OpenAI having needed to release a "baseline" TRPO implementation is a fucking disgrace to your profession?

• Jesus christ, who decided to name a tensor concatenation function cat?