#microformats 2024-10-19
2024-10-19 UTC
[Tim_Nolte], gRegor, lkajfla and [aciccarello] joined the channel
# 
[tantek] [snarfed] trwnh I forgot a great real world example of h-entry for dated pages: W3C technical reports! (specs, notes, statements etc.) E.g. check this out: https://pin13.net/mf2/?url=https://www.w3.org/TR/w3c-vision/
# 
trwnh [tantek] ah yeah i was thinking of undated pages though. i think the data model i was assuming was that h-entry was for that kinda "episodic or datestamped" bit that got removed. this sort of relates to our earlier discussion in #indieweb-dev about h-cite and whether it is appropriate for quotations that aren't also citations (ie there is partial content but not always a cite/source and only maybe attributed to an author)
# trwnh from a generic viewpoint i'm trying to make it so that parsing the web page for mf2 should allow extracting a name (not always the same as <title> or <h1>), summary (which <p> on the page is the summary exactly?), content (it's not always just <main>), author (possibly handled by rel=author but how do you express other properties of the author), and also various relations like in-reply-to (for which rel=in-reply-to is superseded). it looks like
# trwnh h-entry mostly works but i was led to reconsider by https://microformats.org/wiki/h-entry saying "if there is no "published" date for the "entry", then reconsider whether it is correct (or worth) marking it up as an entry."
# trwnh interestingly https://indiewebify.me/validate-h-entry has some text that seems to make no sense to me -- it says "The nested in-reply-to microformat should be an h-cite as it refers to off-site content." even though it's on-site content. but the reason i say it's interesting is that it hints toward the fact that at some point, someone clearly made a distinction between h-entry being "on-site" and h-cite being "off-site". which is an implication
# trwnh that doesn't seem to hold up... does it? in any case, this whole thing has generally been throwing my head for a loop because the types being expressed here don't initially line up with my understanding of type theory (for example, h-entries suddenly becoming h-cites when you refer to them implies that the reference is different from the thing itself -- kind of like how you can reify a Link and then describe properties of it like `rel` or `href`
# trwnh . actually i guess that saying h-entry is an Object and h-cite is a Link is not a bad way of looking at it, now that i think more about it...)
# trwnh er, forgot to specify, i'm using AS2 vocab for that parenthetical at the end, but it applies more generally than just AS2. like as a data model thing
# trwnh in any case i guess the conclusion i'm led to draw here is that h-entry and h-cite started out narrowly defined but have since evolved to cover more use cases than initially designed for
# trwnh https://microformats.org/wiki/h-entry at least the "bare minimum" properties should note gRegor's insight above that "Consumers should generally be prepared to handle no dates, or mangled dates"
# trwnh also https://microformats.org/wiki/h-cite where it says "for publishing citations and references" should maybe have some consideration for evolved usage where an h-cite is a reference more than it is a citation? for example .u-in-reply-to.h-cite as recommended in various places. but i'm still a little wary of so much of the text of the wiki article saying that it's narrowly scoped to citations. i'm also still mostly wondering how to
# trwnh differentiate between mla-style "works cited" and apa-style "references" when processing h-cite, vs something more like a bibliography which i don't want to show up in my list of references. (i am planning to parse my own html for h-cite and auto-generate a references section, to be clear about use-case.)
[jeremycherfas], [snarfed], lkajfla, btrem, sebbu, [KevinMarks], aagaming, barnaby and gRegor joined the channel