#dev 2022-09-12

2022-09-12 UTC
angelo, [tonz]1, Jamietanna1 and gRegor joined the channel
#
@as790726 透過 webmention 來收集 blog 的社群迴響 https://jason-memo.dev/posts/webmention/ (twitter.com/_/status/1569134516098441216)
#
[tantek]4 !march
#
Kaja It's March 926, 2020 (UTC)
#
[tw2113_Slack_] teehee
jacky and bterry joined the channel
#
[jgarber] Went a little wild with the wiki leaderboard scripting and am now grabbing _all_ wiki users and attempting to parse their homepages for microformats2 data:
#
[jgarber] https://github.com/jgarber623/indieweb-wiki-leaderboard/tree/main/data/users
#
[jgarber] ~2750 users, `n` of which are spam-ish/not URLs/don’t resolve properly. Turns out I’m overloading my own service (https://micromicro.cc) by blasting it with requests to parse that many URLs. 🙃
jacky, bterry and [Jamie_Tanna] joined the channel
#
[Jamie_Tanna] Nice, that'd explain the intermittent 500s I'd been seeing ☺ good fix, aaronpk!
nathan[m], zack[m], mambang[m], AramZ-S[m], Steve[m], Pyroxtheythem[m], h4kor[m], neceve and tetov-irc joined the channel
#
@as790726 ↩️ webmention 真的蠻有趣的，但 implement 全部弄完真的要費一番工夫 (twitter.com/_/status/1569282362151948288)
#
capjamesg [snarfed] Did you have any reservations open-sourcing the IndieMap data?
#
capjamesg I'm thinking about doing this for IndieWeb Search.
jacky joined the channel
#
[snarfed] capjamesg no big reservations, no
#
[snarfed] I didn't think about it as open sourcing as much as just republishing, and that they explicitly retain their original copyright etc. https://github.com/snarfed/indie-map/#indie-map
#
[snarfed] I have gotten one removal request so far, which I complied with. https://github.com/snarfed/indie-map/issues/2
#
Loqi [csarven] #2 Request to omit all statements on csarven.ca
#
[KevinMarks] https://xeiaso.net/talks/how-my-website-works
[manton] joined the channel
#
[manton] I love that title. 🙂 Skimming through, nice to see they’re using JSON Feed extensions for some extra metadata.
#
[schmarty] A great site and blog in general!
[Murray] joined the channel
#
[Murray] Having a discussion elsewhere around whether using microformats as CSS hooks is a good/bad idea. I vaguely remember this coming up here a few times, but can't find any resources. Does anyone know any Wiki pages or blog posts that sum up arguments for/against?
#
aaronpk i believe the consensus is it's a bad idea
#
[Murray] that is what I felt, but I'm struggling to explain _why_ that is the case 😅
#
aaronpk the reason i stopped doing it is so that i can change the structure of the microformats classes without affecting the presentation
#
aaronpk i guess that hasn't mattered much lately but was very helpful as i was developing the site especially when adding new microformats markup to it
#
[Murray] that makes sense
#
GWG Same here
AramZS joined the channel
#
@simbafs 剛剛看到一個酷東西，用網站當作登入的入口（應該是這樣說吧 http://indieAuth.com (twitter.com/_/status/1569361423566012417)
#
[tantek]4 [Murray] , aaronpk, pretty sure it's an FAQ (why not to use microformats class names for CSS rules)
#
aaronpk hm not seeing anything specific on https://microformats.org/wiki/faq
#
[tantek]4 hmm, it looks like either it's not or there's possibly some out of date thinking on this 😬
#
[tantek]4 yeah I need to fix this: https://microformats.org/wiki/faq#Q._How_does_the_use_of_class_values_for_semantics_inter[…]with_the_use_of_class_values_for_attaching_CSS_styles.3F
#
[tantek]4 ^ I just updated this FAQ. aaronpk, [Murray] can you review ?
#
[schmarty] Somewhat related, Adrian Roselli recommends using CSS selectors to try and enforce proper use of aria https://adrianroselli.com/2021/06/using-css-to-enforce-accessibility.html
AramZS joined the channel
#
[tantek]4 Fascinating. Will have to read that to see if the reasoning is any different from when we originally thought it was good to use microformats classes as styling hooks as well
#
[tantek]4 Ah that's very different reasoning and use-cases! The stuff about using ARIA as hooks for the state of an element make complete sense, though I feel there should've been at least a mention of the respective CSS pseudo-classes, their similarities, and when to use one or the other or both
jacky joined the channel
#
[schmarty] ah, interesting point. if i understand this advice correctly, wherever a CSS pseudo-class exists typically you wouldn't be using ARIA anyway, since the preference is to use HTML-native elements and semantics first, with ARIA as a fallback.
#
[tantek]4 Yes that is how it's supposed to work. As with anything there's always the danger of overuse (remember when people used to put u-photo on every img tag?)
zack[m] and jacky joined the channel
#
GWG [tantek]4: I think I need to fix some code that still overscopes on that
jacky, AramZS, nertzy and jacky__ joined the channel
#
capjamesg A CSV with a few million links from IndieWeb Search is available now: https://github.com/capjamesg/indieweb-search-links/blob/main/data/2022-09-12.csv.gz
#
angelo nice, i see 1.4 million lines.. i'm going to feed unique domains back into indieweb.rocks
#
capjamesg I crawled your people.txt list plus a few others. I wouldn't recommend crawling the "target" links.
#
capjamesg angelo I can send you raw HTML if it's helpful too.
#
angelo you have 1.4 million pages of raw HTML?
#
capjamesg I have 250k.
#
capjamesg Each page can have multiple outgoing links which is why there are 1.4 million lines of links.
#
angelo if there's a reasonable way to transfer them i could put them to use.. what size are we talking about?
#
angelo if you could limit it to the domain's homepage only, that'd be ideal
#
capjamesg 4.3gb according to Elasticsearch.
#
capjamesg Here's what I have in addition to the HTML: https://github.com/capjamesg/indieweb-search/blob/main/crawler/save_record.py#L198
#
capjamesg (if there's any stats people are interested in, let me know so I can add it to the crawler - crawls run weekly now)
#
angelo what are the h_card and is_homepage records?
#
Loqi It looks like we don't have a page for "h_card and is_homepage records" yet. Would you like to create it? (Or just say "h_card and is_homepage records is ____", a sentence describing the term)
#
capjamesg is_homepage tracks if a page is a root.
#
capjamesg h-card is the h-card found on the page.
#
capjamesg I use that to show profile pictures in search results.
#
angelo is the h-card representative?
#
capjamesg Perhaps.
#
capjamesg Let me check.
#
capjamesg I'm sorry but it's not.
#
capjamesg It looks for a h-card on the page and if it can't find one it does authorship discovery.
#
angelo that's cool. if you could give me a list of the domains with h-cards on them that'd be exactly what i'm looking for. how difficult would that be to pull out?
#
capjamesg Not too difficult. I can share it tomorrow.
#
angelo awesome
#
[tantek]4 plurality << https://twitter.com/tmcw/status/1568966699436621825
#
Loqi ok, I added "https://twitter.com/tmcw/status/1568966699436621825" to the "See Also" section of /plurality https://indieweb.org/wiki/index.php?diff=83413&oldid=51683
jacky, tetov-irc, gRegor and AramZS joined the channel