> > after few months I can't update to a new version
> forum owners often won't upgrade their website ... plugins that likely will break
Maybe then it'd make sense if I mentioned Talkyard https://github.com/debiki/talkyard — it's forum software with automatic upgrades. I'm developing it. There's not yet any plugin system, instead currently "everything" is built-in, and there are (unfortunately) fewer features.
What do you plan to do about the plugin problem? A lot of forums come to rely on these plugins and often times the owners aren't programmers. It means that if they are stuck choosing between upgrading the core software and keeping a useful plugin they often will choose the latter.
Will an automatic upgrade break plugins then? Because I can't see that being optimal.
Maybe one answer can be 1) a public automatic tests repository, 2) release channels, 3) incompatibility auto warnings, and 4) organization internal tests.
1: (for plugin developers) The plugin developers could write automatic tests of their plugins, save them in a plugin-auto-tests source code repository, which the Talkyard developers then would run as part of the Continuous Integration builds (which happens many times more often than new version releases).
The plugin developers would get notified about plugins getting broken, before new Talkyard versions had been released. And could either fix those problems, or tell the Talkyard developers to "please don't do that change at least not now".
2: (for plugin users) Talkyard has release channels (inspired by Kubernetes [1]) — there's a Rapid, a Regular and a Stable channel. (Currently only a Regular channel.) One's Talkyard site could stay on the Stable channel, meaning, there'd be about half a year between the day when a plugin problem got noticed, and a new Talkyard version had been promoted to the Stable channel. That ought to be enough time to have done something about plugin breakage. (Security bug fixes, though, would get added to the Stable release channel, as soon as sensible.)
3: (for both developers and users) There's an automatic notices system in Talkyard. If a plugin uses deprecated functionality, Talkyard can show a warning to the Talkyard forum admins, and then they can decide what to do. Maybe contact the plugin developers? If they don't reply in a month, try finding a new plugin? Contact Talkyard and ask them to take over mainenance of the plugin? (These incompatibility auto warnings are less-noise than subscribing to new-releases-announcement emails, because these warnings would inform you about only the deprecated functionality the plugins on your site use.)
4: If it's super important that a plugin won't break, then the organization could setup a test environment with at a Rapid release channel version of Talkyard. Then they could do their own internal private tests of new versions, about half a year before they'd appear in the Stable channel.
If all these (at least 1-2-3) were in place, I'd think that theoretically, yes plugins could still break. But maybe it would no longer be the right thing to "worry" about. (Auto upgrades are optional, b.t.w., and enabled by default.)
> forum owners often won't upgrade their website ... plugins that likely will break
Maybe then it'd make sense if I mentioned Talkyard https://github.com/debiki/talkyard — it's forum software with automatic upgrades. I'm developing it. There's not yet any plugin system, instead currently "everything" is built-in, and there are (unfortunately) fewer features.