A Discussion in Lion's Internet Office, #wikis
- date: 2021-03-24
- participants: @tme5, @lion
tme5 brings up:
- maintaining page lists manually is a problem – (lion: I don’t understand the reasoning used here)
- the problem that mentioning a category is the same as joining one – (resolving category participation by full text search)
- “or maybe there can be a special way to mark a link as a “strong” link which comes up in a different listing than the usual backlinks”
- ex: PMWiki has a (😝agelist … 😊 special directive, that has options and generates lists
- (Lion suggests potential application of HalfVisualEditor concept.)
Lion brings up:
- if trees are used in wiki, they must support aliasing – the same node appearing in multiple trees
- trees should be rooted in time, because organizing in trees is fundamentally rooted in time (categorizations and categorization schemes are time-specific) – “This is the categorization system that made sense in 2021.”
- tree mechanisms in wiki could be based in textually editable pages, but interacted with graphically, to take advantage of peer review on RecentChanges.
- tags - same thing; All organization of information is time-based, first and foremost. So it’s not just tag “foo”, it’s tag “foo, created and assigned 2021.”
- Lion: “Hm, I don’t quite get what you’re trying to do… You’re trying to make a clear top-down picture of the wiki?”
- tme5: “yes, if it makes sense - but if it doesn’t, a clear picture of some sub-part of a wiki, where it is important that things are complete. I would think most looser, web-like wikis would still find use from this”
Echoes of the ZettelKasten conversation; web-like arrangements of local tree-type structures.
Alex Schroeder: The observation that all organization of information is time-based, with tags also being time dependent, is something that is very much on my mind. With my homepage now having nearly twenty years worth of pages on it, I “solve” this particular issue by searching both for a tag and a page name match, which works since I use a DatePage format. I can search for `#rpg` and filter for `2006` in order to get the slice I want. It feels very crude but it’s a start. I’ve also started investing more into the sorting of search results, in the same crude fashion. When searching for something, the results are sorted first by date descending (to get recent blog posts), and then alphabetically ascending (for regular wiki pages), with comments and images belonging to a date page sorted right after the date page itself.
I remain very interested in the Bliki idea, and the irresistible draw. Perhaps the temporal nature of blogs is more important than I was willing to give it credit for back in 2003. 😊
Other links that might be relevant:
I’m not sure how this is going to help, but at least the tech side of this is solved to a large degree. What’s lacking is adoption. I suspect it’s just not useful enough.