For now this way of publishing is probably only something for true Logseq believers - there are still many, many problems (see how I attempted to fix/work around some of them: https://lamentable.technology/logseq/#/page/website), but I can see it becoming really great with some polishing work. It's definitely cool how easy and fast it is to share bits of knowledge.
I have a fairly detailed idea of how many of the issues I see could be tackled, but haven't gotten around to making a nice post on the forum yet (until then see my note: https://lamentable.technology/logseq/#/page/logseq%20ux%20cr...) or do any actual implementation work (not getting along with Clojure very well). I do have some amount of insight into how my ideas might be implementable though, so hmu if you're interested.
I feel the exact same way. The built in publishing feature is a little rough and not optimized at all. I get the impression that logseq is investing a lot more energy on their own hosted solution than the self hosted publishing right now.
Despite the downsides, as someone who spent way too much time tweaking their own static site, it's really nice to be able to write and organize all the public pages along with my personal stuff inside logseq. I just tag a few notes as public and push them at the end of the day. I was spending way too much time configuring and polishing the static site instead of just getting the notes out there. Seems like a decent option for someone nontechnical to make a site too.
Since Logseq handles everything with plain text, some people are just generating a static site from those files directly, but you lose some of the advanced logseq features.
There's also a plugin to export your notes as a hugo static site. https://github.com/sawhney17/logseq-schrodinger
I'm looking into doing this to make a "light" version of the notes as a static html site at some point, but I'll reevaluate when logseq reaches 1.0 and they solidify their roadmap more.
Your list of issues with the existing system is really nice. I'll link to that in this post when I push my next batch of updates so people are aware of some of those sharp edges.
For now this way of publishing is probably only something for true Logseq believers - there are still many, many problems (see how I attempted to fix/work around some of them: https://lamentable.technology/logseq/#/page/website), but I can see it becoming really great with some polishing work. It's definitely cool how easy and fast it is to share bits of knowledge.
I have a fairly detailed idea of how many of the issues I see could be tackled, but haven't gotten around to making a nice post on the forum yet (until then see my note: https://lamentable.technology/logseq/#/page/logseq%20ux%20cr...) or do any actual implementation work (not getting along with Clojure very well). I do have some amount of insight into how my ideas might be implementable though, so hmu if you're interested.