Hacker News Reader

Comment by noname120 | Hacker News Reader

What caused confusion is that you put “written in ~250 lines of idiomatic Python code” in the readme file and didn't mention anywhere that you were in fact using a bunch of C libraries to do the heavy work.

The disclaimer section (which contains this information) was added 44 minutes ago. And it's still misleading. For example about hiredis you write:

> Python extension that wraps protocol parsing code in hiredis.

Where is the mention that the parsing code is actually C code?

------

Last but not least you already posted this a year ago on Reddit[1] and Hacker News[2] (edit: someone else did for HN), but only (partially) addressed criticism very recently.

[1] https://www.reddit.com/r/Python/comments/awav6k/pydis_a_redi...

[2] https://news.ycombinator.com/item?id=19287717

View on HN

I don't like your tone; this was a one-off experiment (that did not even gain much attention back then) that I did for fun, and no one is entitled to my time so I have addressed those issues when I wanted to.

I am not the one who posted here, why so aggressive?

Fair point, here is a pull request that addresses my criticism: https://github.com/boramalper/pydis/pull/15

> as long as the heavy work is done by calling native non-Python code.

Do not merge that. I once realized that grep is really fast, and replaced our custom solution with bash script, but I still claim that app is written in idiomatic PHP/python/ruby. This project has a lot of problems that it depends on C code is only a symptom. Noting all those bad things in a project Readme is a trap I fall into way to often it never works.

I don't think the tone was hostile and if you ignore any hint of tone you might hear, you have to admit that the title is really click-baity (yes I understand that you didn't post it). Regardless of the criticisms of the project, it's great that you're experimenting and I would be curious to know what's causing the reduction in performance ... is it only part of the program or is that the performance of the Python code being interpreted.

> I would be curious to know what's causing the reduction in performance ... is it only part of the program or is that the performance of the Python code being interpreted.

Mostly the latter but also a bit of the former: I could have optimised it further — at the expense of idiomaticity — but I believe gains would be marginal. Other than that, Python is not that fast so you pay some penalties there too.

Don't listen to these mindless code golfers. Your experiment does indeed prove your point imho.

Reply on news.ycombinator.com