Hacker News Reader

Comment by user8501 | Hacker News Reader

That’s a very good point. I feel, however, that the main idea is that bad philosophy is what ultimately fuels complexity.

The grug brain philosophy is simplicity at all costs, unless absolutely unavoidable.

The big brain philosophy as grug sees it is reusability at all costs, unless absolutely unavoidable.

The issue with this philosophy is that it tends to lock in first generation design choices and makes iteration more difficult.

It’s true that simplicity is difficult to find, but iteration is the key to finding simplicity.

View on HN

> The issue with this philosophy is that it tends to lock in first generation design choices…

I agree.

A common approach to simple is “just start with the first thing that pops into your head and see how far you get.” I guess we could describe that as “simple to think of.”

A far less common approach to simple is “think through the whole problem, then remove everything that you don’t need.” This is what I’d call “a simple solution”, but note that it takes a lot more work to find it.

You've made one strawman accusation, and sure, people polarized that specific way (insisting only things worth doing are worth making reusable) are often doing bad things.

But I see plenty of anti-intectual anti-whatever attitudes founded in other forms of disdain. And I think grug reflects a broad/broader spectrum of negative biases.

One example, I've had an incredibly hard time getting folks to switch from dirty cobbled-together-with-StackOverflow shell scripts to Ansible, which is a just more sight-readable consistent experience. Or to zx or anything not just badly written shell scripts no one collaborates on.

People swear worse is better, are YAGNI (you ain't gonna need it) up the wazoo against 90% of everything. I think grug picks a lot of things to bash on, and a lot of devs do to.

I adore simplicity, but I find it challenging & long & arduous to hammer out of things. It's not fast to produce, or low thought. It's not made by resolutely sticking to lo-fi paths & mentalities.

We do need to be aware of the hazards at the other side, of ridiculous complexity & absurd/unnecessary systems engineering: yes! And I think grug offers some good almost-koans to reflect on, illuminates real hazards well. But I also think it's actively harmful & pernicious to put grug-brainedness on a pedestal, to go about actively disbelieving against any possibilities that might possibly be done simpler.

We seem to agree that simplicity takes iteration. Grug presents it as "oh I'm just dumb grug brain, I dunno" but in practice there is a lot of infighting in tech, and those who insist too loudly on aiming too high are not even IMHO as dangerous as those too loudly insisting on too low. We have to keep engaging possibility from all sides, & keep finding out what balances do work.

> One example, I've had an incredibly hard time getting folks to switch from dirty cobbled-together-with-StackOverflow shell scripts to Ansible, which is a just more sight-readable consistent experience.

Ansible can be used quite elegantly/simply, but the ecosystem as a whole is totally infested with "big brains" who insist on complicating things to make them more Reusable™.

There are really guys out there who will argue that your 20 line shell script actually needs to be two Ansible Roles each consisting of 10 files spread out among 2 layers of directories, and can be published to the Galaxy.

(In reality you can turn that 20 line shell script into a 30 line idempotent, single-file Ansible playbook, which is SOMETIMES worth it)

Reply on news.ycombinator.com