Last weekend, at WordCamp US, I asked a question at the town hall.
At its heart, the question was that did Matt feel the rapid release cycle of WordPress major revisions was too fast? I asked this based on the concerns I hear voiced by plugin developers, generally the smaller shops (single person or under ten teams), who not only have to test their plugins with the new versions of WordPress, but also learn these new, rapidly evolving and changing, libraries like ReactJS and the REST API.
Matt’s answer was essentially no, and that he felt that things would only get faster. Also he said he didn’t think plugins should be one-person shops.
What did you think of the answer?
I was asked variants of this by many people that night and the next day.
I disagree with Matt somewhat.
This isn’t a shock. I’m sure he’ll read this and nod. But he and I both know that a health disagreement can be good for an ecosystem. I understand his point, and in many ways agree with it. A team project for plugins and really any development is what makes things improve faster. Two heads are better than one.
But at the same time, we look back on things like Yoast SEO, and to think that those can only exist while supported by a team is to forget the way that all of this, even WordPress, started.
One person has an idea. One person shares the idea. A second person makes a suggestion.
One person.
Of the 45,000 plugins in the repository, the majority happened because of one person. One person had an idea and a dream and built a plugin. One person learned a thing and shared it. One. And the harder we make it for that one person to grow, the harder it will be for them to become the next Yoast, or Woo, or Jetpack.
As of this post, we released four major releases of WordPress within 355 days. I think that speed for the sake of speed is as bad as dragging out feet and having one release a year. Yes, we have improved our stability by having more frequent releases, because we don’t rush to have an unready feature added to a release. There’s going to be another soon. And that’s a good thing.
At the same time, it’s pushed us to release faster and faster, and that causes the bar to be set too high to new people. It causes burnt out. It causes update fatigue.
I don’t think we should revert to ‘release when it’s ready’ again. That has as many problems (if not more) as ‘release X times per year.’ I do feel we need to consider the emotional health and the supportability of what we are releasing.
Do it well, do it fast, do it cheap. Pick two. And know that the price is from our blood and bone.
I think we should turn it down a notch. Just one notch. And we should stop releasing just to be sure we release a number of times a year.