And if Canonical gets out of line, there’s already been proof of concepts of third-party snap repositories, so that’s a moot point.
That’s not really a moot point. A proof of concept is much different than an implementable specification. And especially in the commercial space, you’d want to host your own snap repo just for that handful of software you want to customize and then reference upstream for the other things. Snaps, to be useful to that user, need to have the same sort of design considerations apt/deb an yum|dnf|zypper/rpm have about multiple repositories and an implicit promise to not break the 3rd party stores version to version.
Additionally an open source server component was a day one promise from Canonical about Snaps (along with monetization and a functional license model for closed source software) and the gas lighting from Ubuntu about why that stuff isn’t here and doesn’t appear to be close gets harder to ignore every year it goes on.
That’s not really a moot point. A proof of concept is much different than an implementable specification. And especially in the commercial space, you’d want to host your own snap repo just for that handful of software you want to customize and then reference upstream for the other things. Snaps, to be useful to that user, need to have the same sort of design considerations apt/deb an yum|dnf|zypper/rpm have about multiple repositories and an implicit promise to not break the 3rd party stores version to version.
Additionally an open source server component was a day one promise from Canonical about Snaps (along with monetization and a functional license model for closed source software) and the gas lighting from Ubuntu about why that stuff isn’t here and doesn’t appear to be close gets harder to ignore every year it goes on.