400,000 GitHub repositories, 1b. files, 14 TB of code: Spaces or Tabs?

Spaces always. I just always program vimrc or editor's config to auto insert a certain number of spaces divisible to some number like 2 or 4.

IMHO thought spaces were always recommended to ensure formatting is consistent across multiple environments so never read too much into the spaces vs tabs on SV TV show.

A lot of coding guidelines like Python recommend spaces.

Next on SV, Vim vs EMacs? ;)

Sorry this is a dup: https://news.ycombinator.com/item?id=12397246

Didn't realize.

I use spaces, but by pressing tab and having it auto-indent the spaces for me.

Never would I waste the energy in my fingers by typing spaces on each line. If I had to do that I'd never use spaces.

Kudos to analyzing the data! I wonder how the duplicate file detection was implemented.

Was it just take a hash of the file contents and see if the hash already exist? Was it compare the file name?

I imagine that many duplicates may exist that contain one file written in one editor and then loaded to another editor, auto formatted (to convert spaces to tabs or vice versa) and then pushed back up. So you might have two files with identical code but different whitespaces.

I also wonder (for Javascript) if the minification steps screws up the results too, since minifiers may convert tabs to spaces or vice versa...

Anyway, I'm probably overthinking this. Great job!