#microformats 2021-04-07

2021-04-07 UTC
KartikPrabhu, [tw2113_Slack_] and [tantek] joined the channel
#
@westurner ↩️ #extruct is a Python library for extracting all sorts of structured data from HTML, RDFa, Microdata, Microformats, JSON-LD: https://twitter.com/westurner/status/1285747050911543303 (twitter.com/_/status/1379592038413373441)
KartikPrabhu, [scojjac], [tw2113_Slack_], iwaim_, lahacker, hendursa1, tomlarkworthy and [KevinMarks] joined the channel
#
Zegnat I am not sure the mf2 parsing spec defines an expected outcome of mf2 inside svgs embedded in HTML. But I would not be surprised if some parsers support it
[arush] joined the channel
#
[KevinMarks] In general svg elements can have classes too, so p- and e- should work but u- properties might need attention and not sure about dt-
#
Zegnat It is more that I think HTML5 spec does not define how elements within <svg> are parsed. So there is no guarantee they are even part of the domtree that mf2 is looking at
#
sknebel it does
#
sknebel SVG and MathML elements do get some consideration in the HTML5 parsing rules
[tantek], [asuh], [scojjac], Phae, [KevinMarks], [manton], hendursa1, tomlarkworthy, [fluffy], lahacker, [email096], [chrisaldrich], [aciccarello], KartikPrabhu, PauloPinto, [tw2113_Slack_], [schmarty] and [jeremycherfas] joined the channel
#
tantek edited /RelMeAuth (+256) "/* detailed algorithm */ check for IndieAuth before performing rel-me discovery (because no one implemented checking a user's own site for OAuth, also OAuth lacked sufficient discovery mechanisms to implement that), explicit should reject silo profiles, may accept 1-2 for testing/demo purposes" (view diff)
#
[tantek] aaronpk ^ switched RelMeAuth to require IndieAuth discovery on personal sites rather than "OAuth discovery" which was never well defined (at all?) on its own anyway
#
[tantek] dropped something that was not implemented (not implementable) in preference for something that *is* implemented
#
aaronpk OAuth discovery was finally published in 2018, but as far as i have seen is mainly only implemented in enterprise use cases, not even consumer APIs like GitHub https://tools.ietf.org/html/rfc8414
maxwelljoslyn joined the channel
#
[tantek] yeah, not for actual open interop, like actual open discovery
btrem joined the channel
#
btrem https://chat.indieweb.org/microformats/2021-04-07#t1617801546277600
#
btrem No doubt I'm stating what is already known, but there is no datetime attribute for any SVG elements, so yeah dt- would be a problem.
#
aaronpk not really a problem, just not needed
#
btrem But then, I'm hard-pressed to think of a reason for using anything other than the `<time>` element for a dt property.
#
btrem I'm also hard pressed to think of a way to use MathML in microformats. Nothing comes to mind at the moment.
#
btrem I'd think it would mostly be `<svg><text/>`. I'm using `<text>` for a `p-name` and `<svg>` for the logo. It's conceivable that someone might want to use `<svg>` for a photo property. The solution that jumps to my mind is `e-photo` instead of `u-photo` (which is what I did for the logo, using `e-logo`instead of `u-logo`).
Seirdy joined the channel