I was hoping it would clear things up, as you said yourself:
it’s not unreasonable for a tag page to also just be a regular wiki page
Of course, who knows, maybe you meant something very different with that.
In the case of LessWrong, the issue is of course that we are dealing with at least two content types: wiki pages and posts. They are written in very different authorial style and have very different premises. This means a tag page cannot just list other wiki pages, it also must list posts.
Having a wiki that just tries to repeat all the text from the long-form articles that the site is already filled with would be a terrible waste of time and give totally disproportionate canonicity to random people editing the wiki pages. For most concepts the best reference is not a wiki page, it’s the article that introduced and explained the relevant concept. So a LW wiki page that fails to link to the best LW essay on an article fails by the standard of the LW wiki.
Normal wikis do not try to handle this issue of integrating lots of long-form essays with an authorial voice into their wiki system. Having wiki pages be tags solves that problem. If you have another solution, you can give it to me (and no, integrating the best essays for each concept into the body of the wiki is not feasible, because that requires too much manual labor, as we quickly found out in early iterations of the wiki during user testing).
I am unhappy with a bunch of the implementation details of how wiki-tags ended up being built, but it’s solving a problem that basically has to be solved on LessWrong if you want to actually direct user attention anywhere helpful.
And the basic abstraction here is also not that weird. Categories or list-articles or wiki-data-tables are all abstractions that other wikis have to handle similar issues.
Again, I think the wiki-tag system has a bunch of issues and has not been our priority for quite a while, but I feel pretty good about the core abstractions here, and also, you did literally just start your comment with “it’s pretty normal for some wiki pages to be tags” and to then express confusion when LessWrong follows this very normal pattern, is at least confusing.
(And to be clear, I am not actually interested in continuing this conversation, but it seemed cheap to clarify at least a bit of the design motivations for posterity)
(fyi I haven’t really been following this discussion, but saw this and thought this outlining of the constraints the LW wiki is under & the motivation for the design was interesting!)
Not every wiki page is a tag! Some wiki pages are tags, which I think makes sense. Others are articles optimized to be wiki pages.
That… is actually even more confusing… :(
I was hoping it would clear things up, as you said yourself:
Of course, who knows, maybe you meant something very different with that.
In the case of LessWrong, the issue is of course that we are dealing with at least two content types: wiki pages and posts. They are written in very different authorial style and have very different premises. This means a tag page cannot just list other wiki pages, it also must list posts.
Having a wiki that just tries to repeat all the text from the long-form articles that the site is already filled with would be a terrible waste of time and give totally disproportionate canonicity to random people editing the wiki pages. For most concepts the best reference is not a wiki page, it’s the article that introduced and explained the relevant concept. So a LW wiki page that fails to link to the best LW essay on an article fails by the standard of the LW wiki.
Normal wikis do not try to handle this issue of integrating lots of long-form essays with an authorial voice into their wiki system. Having wiki pages be tags solves that problem. If you have another solution, you can give it to me (and no, integrating the best essays for each concept into the body of the wiki is not feasible, because that requires too much manual labor, as we quickly found out in early iterations of the wiki during user testing).
I am unhappy with a bunch of the implementation details of how wiki-tags ended up being built, but it’s solving a problem that basically has to be solved on LessWrong if you want to actually direct user attention anywhere helpful.
And the basic abstraction here is also not that weird. Categories or list-articles or wiki-data-tables are all abstractions that other wikis have to handle similar issues.
Again, I think the wiki-tag system has a bunch of issues and has not been our priority for quite a while, but I feel pretty good about the core abstractions here, and also, you did literally just start your comment with “it’s pretty normal for some wiki pages to be tags” and to then express confusion when LessWrong follows this very normal pattern, is at least confusing.
(And to be clear, I am not actually interested in continuing this conversation, but it seemed cheap to clarify at least a bit of the design motivations for posterity)
(fyi I haven’t really been following this discussion, but saw this and thought this outlining of the constraints the LW wiki is under & the motivation for the design was interesting!)