#microformats 2022-02-23

2022-02-23 UTC
Seirdy, ur5us, jacky, synkarius, gRegor, gRegorLove_, ur5us_, Johan and sayanarijit[d] joined the channel
#
@kevinmarks ↩️ You can use classes for that, which gives you microformats. (twitter.com/_/status/1496454731337764866)
#
@ade_oshineye ↩️ At that point you’d be fighting the markup language rather than focusing on thinking. The best notations eventually become transparent to the user. HTML with microformats is the opposite of that. (twitter.com/_/status/1496456249814228993)
#
ccx Hmm. I do thoughts about that. Knowledge representation is rather broad topic.
#
ccx There are quite of ways to approach it. One somewhat wild one would be to just take your notes in Lojban. That would have for most statements have representation in Datalog, CyC or similar system.
#
ccx And I did write down stuff as Prolog knowledge-base before. It does somehow fit how I think. But this is more about referencing wide range of different things in a flexible manner than it is about well-formed structured data. More XRI and less Microformats.
#
ccx Notably there's one system that did this kind of linking very well on desktop and that is Plan9 "plumber" (often together with Acme as main interface, but the beauty is that it can work with nearly everything)
#
ccx Specifically related to zettelkasten; here's screenshot of implementation on top of edwood+plan9port: https://l2.re/xFQ5fD
#
ccx But if we are indeed talking about writing stuff like ical/vcard entries in plain readable text I was always fan of reStructuredText interpreted text and directives. The goal of rST, driven by the hard to extend nature of old ST, was to have as much as possible forward compatibility and extension points if we want to put in structured data or other features.
#
ccx Now, rST definitely isn't the prettiest format around. But it was the first popular one to bring this feature as far I know. Pandoc did end up defining it's own flavour of Markdown some time after that with similar feature I believe.
#
ccx It seems to have combination of ::: block delimiter and {foo} inline interpreted text, dunno if those are somewhere formally defined as extension points. I generally prefer the rST syntax despite it's various warts (the wide variety of Sphinx plugins is a plus). Either way thit could easily produce microformat-enriched HTML when HTML output is requested.
#
ccx Oh my, I've missed quite a few words while IRC-blogging. I need more tea before doing that I guess. :D
#
ccx * I do have thoughts about that. * There are quite few of ways… * but the beauty of it is…
KartikPrabhu, [aciccarello], angelo, barnaby, gRegor, gRegorLove_, gRegorLove__ and [tonz] joined the channel
#
gRegor Re: https://chat.indieweb.org/dev/2022-02-23#t1645644863649700 I'm done to aim for a 1.0.0 release!
#
Loqi [barnaby] October this year will be php-mf2’s 10th birthday
#
gRegor *down
#
barnaby cool! I’ll make a milestone for it and write up some of the goals I think would be useful
#
[tantek] it would be nice to get things to the point where we are not expecting any breaking changes in the future, e.g. we knew the img/alt change was coming and would be breaking
#
barnaby [tantek]: you mean the mf2 parsing spec itself?
#
[tantek] other considerations of that sort are id (which we already did add), language, and maybe ARIA attributes as well (as in, what if anything should an mf2 parser retrieve from them and represent in results)
#
[tantek] yes
#
[tantek] both the parsing spec and thus the parser implementations
#
[tantek] since changes to one require the other 🙂
#
[tantek] there's also the longstanding half-joking question of "metaformats" that we should consider
#
barnaby metaformats?
#
[tantek] hah yeah it's been kind of a long-running potential april fools joke that may actually have practical implications
#
barnaby what are metaformats
#
barnaby not documented on the wiki then I guess :P
#
[tantek] Loqi doesn't answer what is in this channel
#
[tantek] that is, metaformats are mf2 being applied to meta tags as a use-case, how to parse them, and then defining backcompat rules for parsing popularly supported meta tags like OGP, Twitter Cards etc. into a consistent model based on metaformats
#
barnaby huh that is interesting
#
[tantek] I thought I'd written something somewhere about this before but it may all (mostly?) be in my head or personal notes
#
[tantek] ah yes I did: https://indieweb.org/metaformats
#
barnaby sounds similar to the mf2-shim thing I was using for a while, but which became impractical to maintain, which would parse silo pages into mf2 structures
#
barnaby so I guess the big question is: should this functionality be part of the main parsing spec and expected to be implemented by generic parsers, or should it be outsourced to external libraries
#
[tantek] external libraries are less likely to be consistent / interoperable so that's why there's benefit to specifying it explicitly
#
[tantek] in particular, answering the questions of how to treat mf2 class names on <meta> tags (right now it's undefined)
#
barnaby sure, that’s definitely a valid question for parsers
#
[tantek] once we have a good answer for that (which I have some ideas on), then it becomes relatively straightforward to redefine OGP/Twitter Cards as a form of backcompat
#
[tantek] similar to mf1 backcompat
#
[tantek] specific terms for specific h-* vocabs
#
barnaby unrelated to this specifically, but about the parsing spec: since I was gone, has there been any discussion about how multi-line plaintext values are parsed? the way parsers currently handle line breaks and img src and alt text in p- and struct value keys can make them hard to use
#
barnaby e.g. with php-mf2, the plaintext values have <br> elements replaced with a line break, but a <p>text</p><h2>header</h2> will results in “textheader” in the plaintext value
#
barnaby although this could be an issue specific to php-mf2
#
aaronpk has some serious PTSD from these discussions :P
#
barnaby heh yeah I bet
#
barnaby are the discussions, or at least the conclusions from them, archived somewhere?
#
aaronpk i think there's some discussions on a few open issues on the php parser, maybe also some still open on XRay. some of it was in person convos too
#
sknebel and https://github.com/Zegnat/php-innertext / https://github.com/sknebel/python-innertext as prototypes
#
Loqi [Zegnat] php-innertext: 🏃🐉 Run. Here be dragons.
#
sknebel I think Zegnat made a test site with his that was then linked to some xray issue to discuss variants? vague memories ...
#
barnaby hmm okay Zegnat/php-innertext looks good
#
barnaby and https://html.spec.whatwg.org/multipage/dom.html#dom-innertext
#
sknebel we basically built the two in parallel to check we had a common idea
#
barnaby sknebel: at a glance, both implementations look good to me. did they end up getting used for anything?
#
sknebel no, got to "heres a starting proposal" and then kind of fell asleep
#
barnaby hmm okay
#
barnaby looks like this is the relevant issue https://github.com/microformats/microformats2-parsing/issues/15
#
Loqi [Zegnat] #15 What should mf2 textContent parsing result in? User expectation vs. DOM specification.
gRegorLove_ and ur5us_ joined the channel
#
gRegor There was also https://pin13.net/mf2/whitespace.html but 404 now. Still have that aaronpk or is it not relevant now? Found in 2018 comment in ^
#
aaronpk renamed https://pin13.net/mf2/blankspace.html
#
aaronpk i updated the comment
KartikPrabhu joined the channel
#
gRegor Thanks, updated in my comment too
[chrisaldrich] joined the channel
#
[tantek] redirect to fix the link?
#
aaronpk sometimes i wish nginx had an htaccess equivalent
#
[tantek] HTML meta refresh? 😉
#
aaronpk oh gosh
#
aaronpk ok there you go
gRegorLove_ joined the channel