Snaps are still slow

Don't get me wrong. I have nothing against snaps, contrary to the typical discussions on reddit. But I am actually surprised to see that snaps are still slow.

In some snaps, you can´t noticed the difference; firefox, audacity, to name a few examples. But in most cases, they are slow to the point that you don't see any benefit from using them; despite that I have an actually powerful laptop (i7 10750 h with plenty of ram and ssd).

I wanted to switch libreoffice from deb to snap, since there was a considerable version difference between them; 7.3 --> 7.5. However, after doing successfully the update, loading pages, moving through the files was just awful and I had to go back to the deb format.

Is there any way in which I can report through telemetry the actual performance and the issues? Something like the gnome-info-collect, which by the way, I have it on my machine. I actually would like to see snaps to be a more reliable and successful package format, but right now, deb is most of the cases the best choice.

EDIT 1: Thanks for the responses. I will leave the question open, since this is an ongoing issue, and it might lead (eventually) to a change from Canonical.

I like snaps. I like the idea and what they do. But I do not like where they are right now. What happens with the Libreoffice snap is just an example, and hopefully it will improve.

Many may not like telemetry, for obvious reasons. But I think that as long as it serves the purpose of feedback, to help developers to improve the system that we use. And as long as it is asked, not enforced, it might be a good idea.

mook765 avatar
cn flag provides [appimages](, so you could try this if you want to use a newer version than the deb package. Remind that you need the `libfuse2`-package installed to be able to run appimages, depending on your Ubuntu version, it may be preinstalled or not.
cn flag

"Is there any way in which I can report through telemetry the actual performance and the issues?"

Most snap packages are created by volunteers or upstream projects, so there is no centralized "telemetry" infrastructure. And no guarantee that the author would pay attention to such reporting if it existed.

The preferred way to report a slower snap is to file a bug report with the snap author, carefully detailing your measurements and troubleshooting.

You might check previous bug reports and documentation for that snap -- the author may have already published guidance for measuring and troubleshooting their snap.

I sit in a Tesla and translated this thread with Ai:


Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.