I think requiring an addon is optimal, and I'll explain why.
I see "why can't it just be an option?" a lot. Maintaining an option has a cost. Every option introduced in software creates an additional code path that must be maintained, and must be made to play nice with other features and their options. This can add up rather quickly, especially when options have many subtle interactions with other options. It's not about how easy one option is to maintain, it's about how easy all options are to maintain when we decide to add this one.
In the case of the addon, however, the cost of the option is removed from the main code base and the developers who work in it, and shifted to the addon which is (presumably) maintained by developers who care about the feature. In this way, the cost of the feature is distributed over code and development effort that best serves both the users who care about feature in the addon, and the project's user/developer community as a whole. With the only caveat that the addon api must be sufficient, the addon solution is optimal.
There's a mentality that moving a feature to an addon/plugin/whatever makes it a second-class citizen, but I think in most situations, if executed correctly, it's better than trying to maintain it as an integrated feature.
I see "why can't it just be an option?" a lot. Maintaining an option has a cost. Every option introduced in software creates an additional code path that must be maintained, and must be made to play nice with other features and their options. This can add up rather quickly, especially when options have many subtle interactions with other options. It's not about how easy one option is to maintain, it's about how easy all options are to maintain when we decide to add this one.
In the case of the addon, however, the cost of the option is removed from the main code base and the developers who work in it, and shifted to the addon which is (presumably) maintained by developers who care about the feature. In this way, the cost of the feature is distributed over code and development effort that best serves both the users who care about feature in the addon, and the project's user/developer community as a whole. With the only caveat that the addon api must be sufficient, the addon solution is optimal.
There's a mentality that moving a feature to an addon/plugin/whatever makes it a second-class citizen, but I think in most situations, if executed correctly, it's better than trying to maintain it as an integrated feature.