Hacker News Reader

Comment by userbinator | Hacker News Reader

...showing just how useful "best practices" actually are.

I associate that phrase with dogmatic cargo-cult "don't think just follow" and the horrible results thereof. Lots of self-proclaimed "experts" love to say "do X and Y and Z and you will be successful because these are best practices", but it's all a bunch of snake oil.

https://agilepainrelief.com/notesfromatooluser/2010/03/there...

http://www.satisfice.com/blog/archives/27

http://www.satisfice.com/presentations/nobest.pdf

"Best practices are best not practiced."

View on HN

You probably use every day at least one of the projects with a badge. That includes the Linux kernel, Kubernetes, Node.js, curl, GitLab, LibreOffice, and many others. See a short list here: https://bestpractices.coreinfrastructure.org/en

Or the current complete list here: https://bestpractices.coreinfrastructure.org/en/projects?gte...

As far as being not practiced, that's absurd. We worked to capture the practices already being used by successful open source software projects. Are you seriously arguing that projects should not use Version Control? Are you arguing that projects should not have automated tests? Are you arguing that projects should not have licenses? In short, have you even looked at the criteria?

Also, snake oil salesmen made all sorts of claims because they were going to earn money by selling you a product. Earning a badge requires $0, so the badging process is not a sham to earn money.

You probably use every day at least one of the projects with a badge.

Yes, but it doesn't factor one bit in evaluating whether I use a project.

In short, have you even looked at the criteria?

Yes, and it doesn't matter --- whether or not a project does X is not necessarily an indicator of its success. You're essentially saying "all successful projects do X, so if you do X you'll be successful" --- where success is, in this case, earning a badge.

I've seen far too much software that happens to tick all the "best practices" boxes, yet is a horrific failure at doing what it needs to do.

Also, snake oil salesmen made all sorts of claims because they were going to earn money by selling you a product. Earning a badge requires $0, so the badging process is not a sham to earn money.

It's not free in terms of the time expense.

The best practices badge criteria are focused on increasing the likelihood of developing good software. Neither it nor anything else can guarantee success.

In particular, you can develop high-quality software that does not solve a problem that anyone else wants to solve, in which case it will not be a success.

Guarantees are hard to find. That said, I think it is wise to look at the practices of successful projects and try to copy them where they make sense. That's what this set of criteria does. If you want something that absolutely guarantees success of your project no matter what, then I'm sorry, I don't know where you can get that.

You're right that doing this requires time. In particular, if you don't have a test suite, picking a test framework and using it to create a test Suite takes time. I think it is time well spent.

Best practices are a bit like good genes. By no means a guarantee of success, fame, glory and riches, but damn if they don't make things easier.

I see absolutely nothing dogmatic or cargo cult about the recommendations they make. They are completely sensible, and a decent guideline for improving the technical support infrastructure of a project.

What's wrong with that? Do you have any comparable list of actionable things that projects can do to improve their services?

That resonates. If living in a universe of one then personal best practices are whatever suit your fancy. But in this world apps must interact with a community of apps and be maintained within that context plus without suitable protocols inconsistent results will appear across apps. Some of Kaner, Bach and Pettichord work seems contrarian for its own sake. As in your ego is just as good as purported expert egos

Reply on news.ycombinator.com