On Linus Torvalds, technical & corporate communications

Linus Torvalds has long been one of my heroes. The invention of Linux & the subsequent development of Git were technical and organizational miracles. You could fill a book simply by quoting examples of Linus dissecting technical problems to their components and making it obvious what should happen.

However, over the past decade, Linus’ communication style has degraded from ‘Finnish style robust’ to needlessly hurtful screeds, tearing into people who did not deserve that. For many of us this was painful to see. “Why, Linus, why?”

It now appears that several unnamed community members have breached Linus’ emotional hull and caused him to see the light. This is no mean feat, and let us all hope Linus comes out as a better person. We could all use his insights, especially if delivered in a way that does not hurt.

In expressing my joy at the news of Linus seeking help quite a number of people opined that we should not be “snowflakes”, and that this “Finnish army style communications” were what made Linux great. One of the printable responses:

“No. Just no. You’re so successful because you’re one of few people who don’t waste time beating around the bush. You call a spade a spade instead of polite “professional” bullshit”

And I think this touches on some of the confusion among people that support Linus’ historical combative communication style. They conflate swearing at people with being honest or “straightforward”. And I can see where their confusion is coming from.

Three communication styles

“Hey, we have this moveFile(Fileptr) function, but it needs a flag to not move symlinks. I didn’t want to change the API, but I note that all pointers are divisible by 2. So I propose we use ‘moveFile((Fileptr)((char*)f +1))’ to indicate that if the file is a symlink, it should not be moved. moveFile can then & the pointer to get rid of the 1”.

If you are a C or C++ programmer of any merit, your brain should explode at this suggestion. Here is the “Linus Torvalds” response to this suggestion:

“What the fck is wrong with you? Your employer has been releasing shit products for years, I wonder if this is how that happened. This is like the worst possible API abuse ever and the world will burn down because of this crap. Go away and don’t waste my time with this fcked up shit”.

First the positive. This response is actually what everybody is thinking. It is the unvarnished truth. But that is the only thing this reply has going for it. The author may have had many reasons for proposing this excremental idea and we may not have been aware of all the constraints they were working under. We can safely assume that this person will now become an ex-contributor to our project though.

Contrast this with the ‘modern corporate HR-compliant response’. Bear with me:

“Hi John,

I hope this email finds you well!

Thanks for your suggestion. moveFile() is indeed an important part of our project, and I’m happy to see that you want to improve it. I’ve thought about your suggestion, which is clever in its own way, but I do wonder why you went down that avenue.

The new calling syntax is somewhat unusual (at least to me, perhaps it is idiomatic in other places?) and may lead to surprises. I’m also not entirely sure if we might not one day get unaligned pointers on some platforms. We might get hard to debug problems from that, but I’m not sure, what do you think?

Finally, I wonder how hard it would be to simply extend the moveFile API with a flag ‘noSymlink’, would that be too difficult? It may be there are reasons I’m not aware of, but from where I am sitting, it looks like we could just add that and recompile, but I could be wrong.

Please let me know your thoughts and we can discuss.


🖨️ Please think about the environment before printing this email!”

Wow that was quite a read. Not only did it take a long time to write, it is also a complicated thing to read. The hidden meaning is “no way we are going to do this”, but it spends five paragraphs saying that. You could definitely read this reply wrong and actually go discuss the merits of your odd-pointer idea. I mean, it said this technique was “clever in its own way”! Quite some time has already been wasted on this email and it is pretty likely far more will be wasted in further discussion.

This communication style is almost mandatory in many megacorporations, and it does kill technical progress.

Here’s the “hey we’re all engineers here”-style response:

“Um well - we could solve it this way but it looks fragile and not very obvious to the caller. Is there a good reason we can’t simply add a flag to moveFile?”

In my world, this answer is near perfect for a team that already knows each other. It communicates there is a problem with the suggestion, but leaves open that there might be a good reason for doing it like this. This took 20 seconds to write, and for a typical technical reader, there is no problem interpreting this response.

Note that in many corporate environments, this email message is already worryingly direct and might upset some non-technical people. You would not do this in the marketing department and have a decent career prospective, for example.

If a team does not yet have experience communicating so compactly, slightly more words may be required, or even a private reply that gives some more context that we appreciate the effort, but that technically speaking this won’t fly.

The problem

The problem now is that Linus suggest he wants to move to the “hey we’re all engineers” style of communicating, still making points concisely, but now without swearing or intimidating anyone. However, many people have conflated “not insulting people” with mandatory pages long beating around the bush corporate drivel. And we definitely don’t need that.

I’m charitably going to assume that the people supporting Linus’ legacy communication style are actually worried we’ll soon all have to couch our words in layer upon layer of corporate drivel.

But the good news is that we don’t have to. It is entirely possible to communicate concisely and directly without insulting people or making them angry, and I recommend that any engineering organization strives to talk to each other that way.

And if you have any thoughts, please let me know :-)

Note: my earlier post ‘Talking to technical people’ discusses some similar themes.