Post 124647

@hackernews
Developers who use spaces make more money than those who use tabs
https://stackoverflow.blog/2017/06/15/developers-use-spaces-make-money-use-tabs/
» Comments [news.ycombinator.com]
@hackernews: Developers who use spaces make more money than those who use tabs [stackoverflow.blog]”

There's a long, repetitive comment thread there, mostly just tabs/spaces holy war (die spacer infidels!) but occasionally touching on the actual statistical anomaly. (1/n)
@mcdemarco
The post accounts for the obvious confounding factors using other data from the source SO survey, and encourages the readers to come up with more. Of course they come up with the ones already accounted for...

(2/n)
@mcdemarco
On the actually novel front, the evil spacer infidels leaned towards the idea that spacers are better team players or work for bigger companies -- and thus that the effect was real in some sense. The tabbers didn't have a common theory like that.

(3/n)
@mcdemarco
The best theory was repeated and/or rediscovered many times in the thread, including by posters of no obvious affiliation: people misunderstood the survey question in a way that consistently pushed them into the tabber group, weighing it down. (4/n)
@mcdemarco
The most entertaining part of the thread was people insisting that no one could possibly have misunderstood the survey question, just a few posts up- or downthread from obvious statements of misunderstanding of the survey question. (5/n)
@mcdemarco
I conclude that developers who don't understand a question about tabs vs. spaces know less about development overall than those who understood (or just have a tendency to misunderstand) which leads to lower salaries in a (partly) meritocratic system. (6/6)
@prometheus
@mcdemarco so I will avoid reading the survey;) but note that which language is used bears heavily on productivity and with C/C++ being more dangerous that Python [www.secnetix.de] for people who mingle whitespace glyphs.
@mcdemarco
@prometheus They controlled for language; most of the comments overlooking that involved Go, which may say something about sampling issues when polling SO users.
@mcdemarco
@prometheus One largely overlooked part of the analysis (even in my theory) was that they set aside the smaller group of mixed whitespacers because they weren't significantly different from pure tabbers in outcome.

Go to the App Directory to explore the network more!