My concern is that Lemmy is not scaling and was not tested with enough postings in the database. These “nice to have” slick UI features might have worked when the quantity of postings was much smaller, but it puts a heavy real-time load on the database to search postings that keep growing in table size every day.
I also suggest that this kind of feature be discussed with smartphone app and laternate webapp creators - as it can really busy up a server dong text pattern matches on all prior posting content.
Until a in memory search backend like Meilisearch or Elasticsearch is implemented this seems indeed like a very bad idea. I guess they wanted to do that and got sidetracked?
Another performance concern I have is the tracking of per-user messages being new. I haven’t traced through the code to see how this is being done. Is the database tracking every posting and/or comment that a user is reading, or is this just browser storage? Performance being my concern here if it’s the server doing the work. Anyone know?
I created a community for !lemmy_ui@sh.itjust.works
I would prefer that technical communities about Lemmy be on lemmy.ml, but the creation of new communities is turned off… I know some of the major sies (Behaww, Lemmy.world, SJW) are not federating with each other, but do federate with lemmy.ml - so that community isn’t avilable for all of the Lemmy users.
