Thank you for bringing this up with the team! While it’s unfortunate this has been put on ice for now, based on the discussion post and launchpad bug, it seems that the design choice of migrating ~/snap to ~/.snap/data was welcomed by the community and there was major backlash only to ~/Snap migration. From my understanding it is a conflict between the design choices below:
The Consumer Facing Preference: The desktop should be transparent and manageable for all users. A user’s application data is their data and should be visible and accessible (like ~/Documents), not hidden in a complex structure. Kind of what Apple does on macOS with ~/Library.
The Community Preference: The home directory should be “clean” and free of clutter. Application data belongs in standardized, hidden XDG directories (~/.config, ~/.local/share, ~/.var), and users who need this data are technical enough to find it. This is what flatpaks do and store data in ~/.local/share/flatpak/.
Now, I’m interested in the Canonical devs perspective on the following:
-
Why wasn’t migrating to
~/.snap/datachosen after seeing support for it from the users? This aligns with the community choice instead of~/Snap, is similar to how Flatpaks handle user facing data, and keeps the home directory neat. For example, A consumer first gaming OS like SteamOS is using flatpaks for app support which again stores user facing data in a hidden directory so isn’t something we should rule out completely from a preference stand point because it also seems like the majority users are more accepting of a solution like this. -
If the primary goal is discoverability, why not use
~/Applicationsor~/Applications/Snap? Users are already adopting~/Applicationsfor their AppImages. Leveraging this existing behavior would create a single, understandable “Hey, all the apps are here!” location. This seems like a practical, one-stop solution as well that aligns with the Ubuntu Core desktop goal, providing users a default directory to store and access their Snap data (and AppImages as a nice addition). This was also a suggested on launchpad sometime back #365
I would be very grateful if you could share my questions with the snapd team as well. I value your and the snapd teams’ opinions so hearing back on these questions would help me frame a better proposal. 