I’ve been writing blog posts for more than 20 years at this point, and in the process I’ve learned what is the kind of content that people are more likely to engage with, and which the more likely to withstand the test of time.
One consistent feedback I’ve received throughout the years is that I should use a less combative tone: “be nicer”. According to these critics, if I were more polite more people would be amenable to my message. They say this with all the certainty of a dogmatic belief, but it’s merely a hypothesis, and how does this hypothesis align with the data?
Polar opposites
Let’s look at one of the most scathing posts I’ve written: GNOME’s horrid coding practices.
In this post I criticized very harshly the coding practices of the vte developers, an important component in the GNOME ecosystem, used by many terminals. But it’s not just the awful coding practices that are to blame, it’s also the attitude of developers towards bug reports, and the authoritative nature of the moderators — who threatened to ban me (yet again) just for criticizing the developers.
When I say their practices are horrid, I’m not just being sensationalist: I truly do believe they are horrid.
Stats: views: 9.1 k, comments: 40, likes: 7.
How about a more mild one? Good taste.
This is a completely technical post in which I’m not criticizing anyone, all I’m doing is explaining a notion of “good taste” in code according to me. The idea started with a project by Mark Kirchner (linked-list-good-taste) which I liked very much. I contributed patches at the end of 2020, but eventually I decided that more could be done and created a new project from scratch where Mark’s example would only be part 1 (good-taste). It was also important for me that the result was actual HTML pages, the code examples had syntax highlighting using my own theme, and the illustrations were automatically generated from code.
It took months to finish all 3 sections. I was quite happy with the result, but after I published it I didn’t see any ripples. Was anyone even looking at it?
So I wrote a blog post explaining my motivation behind the project, where it started, and a summary of the code examples with a link to the project.
How did all this work pay off?
Stats: views: 122, comments: 0, likes: 1.
I spent so much time working on my good taste project and in the end only a handful of people saw it. My blistering criticism of vte got more than 75 times the views, and although it took me some effort, it was not nearly as much as this.
Let’s remember that this is a 100% technical post which is not criticizing anyone or making any value judgement about any project.
The evidence suggests that my critics are wrong: people like to see criticism, even if it’s just to criticize the criticism.
Boldness
There’s many examples where a polite tone doesn’t get results, whereas a combative one does. The whole reason civil courts exist is precisely because being nice does not always work.
On the Internet nobody can hear you being subtle.
Linus Torvalds
I wrote a post about a situation in the Ruby community which proves being nice did not work. When I sent purely technical patches to the mailing list ([PATCH 0/2] Fix strptime ‘%s’) absolutely no one replied, but when I sent a fiery criticism, tons of people did: Who the f*heck is Tadayoshi Funaba and why can he reject sensible patches unilaterally?. Sure, a lot of people criticized my tone, but many others agreed with me that there was a technical problem.
Criticize my boldness all you want: it gets results. I got the bug fixed, and my post (My tone doesn’t make me wrong, or how I convinced the Ruby project to fix an inconsistency) was successful.
Stats: views: 1.7 k, comments: 8, likes: 4.
Anyone arguing that being polite is more successful would need to go back in time and make all my attempts of being polite actually successful if they want their argument to have any weight.
The truth is that the power of politeness is a myth, a hypothesis that is contradicted by the evidence in the real world, and only held by purely dogmatic reasons.
Top controversies
Do you think I cherry picked some isolated examples? Nope.
- After two weeks of using GNOME 3, I officially hate it (55 k)
- The problem with GNOME 3 (43 k)
- No, mercurial branches are still not better than git ones; response to jhw’s More On Mercurial vs. Git (with Graphs!) (29 k)
- The Linux way: never ever break user experience (13 k)
- Is ‘git pull’ broken? If so, what’s the fix? (12 k)
Some might wrongly conclude that I’m advocating to always be aggressive. That’s not true. My top post is fairly benign (Advanced Git concepts: the upstream tracking branch).
Can politeness work? Absolutely. But not always. There are cases where a little controversy helps, in this blog I don’t shy away from that, and clearly a good amount of people like that.
Felipe Contreras 