In this episode, John James Jacoby and I start off by providing updates on some items from last week’s show. We discuss WordPress.com’s privacy policy change, the new detailed activity log in the WordPress for iOS app, and cPanel’s acquisition. We end the show with a thoughtful conversation about forking software.
Stories Discussed:
WordPress.com Boots Sandy Hook Conspiracy Theory Sites, Bans Malicious Publication of Unauthorized Images of Minors
WordPress for iOS 10.6 Adds A Detailed Site Activity Log
Oakley Capital to Acquire cPanel
Drupal.org Migrates Developer Tools to GitLab
ClassicPress: Gutenberg Not Included
WPWeekly Meta:
Next Episode: Wednesday, August 29th 3:00 P.M. Eastern
Subscribe to WordPress Weekly via Itunes
Subscribe to WordPress Weekly via RSS
Subscribe to WordPress Weekly via Stitcher Radio
Subscribe to WordPress Weekly via Google Play
Listen To Episode #328:
[powerpress]
By and large, I think there is some reasonable discussion there about merging things into core early. The “plugins as features” paradigm has been tried for quite a while now, and while it is resulting in more “complete” pieces, it also seems like it is taking those longer and longer times before they are ready.
Don’t get me wrong, I can see the benefits of building these big pieces of functionality separately, but at the same time, the audience for those pieces is much reduced by not having them in core very early on. When you are insulated and working amongst yourselves in a small group, then yes, you can iterate quickly, but then as you get further and further away from the main project, then you end up being isolated and alienated from it.
Maybe we should consider for any new features like this to halve the distance. Use the feature-as-plugin paradigm up until to have a workable product, regardless of completeness, then merge it immediately into trunk. Get it out there, bring more eyeballs on it instantly. Trunk gets a LOT of views, and a lot of people run it. And the kind of people who run trunk don’t generally mind brokenness or incompleteness. You can iterate rapidly, but always with the notion of merge-to-trunk is a couple weeks away. Keep that pressure on.
Maybe then we can get the most usefulness and speed out of this idea of features being separated from the main line. Because we always get value out of the main line, separating everything important leads to a vast gap, which seems undesirable. In the long run.