#dev 2022-11-07

2022-11-07 UTC
barnaby joined the channel
#
[tantek]1 design << Useful design considerations for any software/service you build for more than yourself: 2017-05-10 [https://medium.com/a-change-is-coming/lessons-from-mastodon-for-independent-social-networks-ae2d4ccf8f72 Lessons (so far) from Mastodon for independent social networks]
#
Loqi ok, I added "Useful design considerations for any software/service you build for more than yourself: 2017-05-10 [https://medium.com/a-change-is-coming/lessons-from-mastodon-for-independent-social-networks-ae2d4ccf8f72 Lessons (so far) from Mastodon for independent social networks]" to the "See Also" section of /design https://indieweb.org/wiki/index.php?diff=84251&oldid=84080
jacky joined the channel
#
[tantek]1 [snarfed] I think a dynamic allow list of: "current followings" + (set of all @-names / domains I've mentioned in all past posts) - (explicit block list)
#
jacky a "previously engaged with" list, interesting
#
jacky I can see that being sorted/truncated on particular properties (for me, how often I do I tend to reply to them, do they appear a lot o
#
jacky *appear a lot in other specified feeds (for example, crawling /chat-names)
#
[tantek]1 hey who is recently publishing u-follow-of posts? e.g. to send AP follows via Bridgy Fed?
#
[tantek]1 could you add u-follow-of to https://microformats.org/wiki/h-entry#Proposed_Additions and file a GH issue about it (including current known publishers and consuming code/services) https://github.com/microformats/h-entry/issues ?
#
jacky makes a note to do so when he does
#
jacky what is search
#
Loqi search in the IndieWeb usually refers to searching your personal site for your own content (and/or caches of content you’ve responded to), sometimes searching IndieWeb chat archives or the IndieWeb wiki, or the nascent IndieWeb Search index and service to search across community posts https://indieweb.org/search
[dmitshur] joined the channel
#
[dmitshur] [aaronpk] Thanks for linking to https://blog.joinmastodon.org/2018/06/how-to-implement-a-basic-activitypub-server/ from a page on your site. It ends with "But there is a lot we haven’t covered: How to follow ..." which I might try to learn more about next.
#
IWDiscordGateway <kongaloosh> hey, one more dumb question: I've got a list of u-follow-of entries, and I sent an example mention to BF. I'm getting a 501 that says the action isn't implemented. It looks like the errors are coming from here:https://github.com/snarfed/bridgy-fed/blob/001cd1fe11b8333e449a711b6220b8e04adf3f81/activitypub.py#L140
#
IWDiscordGateway <kongaloosh> am I barking up the wrong tree for sending follows? It seems Accept and Create are the only ones allowed.
#
IWDiscordGateway <kongaloosh> I suppose it should technically be a create 🤔
#
IWDiscordGateway <kongaloosh> so it's probably moi
#
IWDiscordGateway <kongaloosh> I probably borked my h-entry spec
#
IWDiscordGateway <kongaloosh> Ahhh, I think it's my u-follow-of.
#
[tantek]1 does the IndieWebify.me validator help you check your h-entry?
#
IWDiscordGateway <kongaloosh> I think the issue was that I wasn't sure what I was supposed to put in the u-follow-of and had the wrong link, tantek
#
IWDiscordGateway <kongaloosh> but I've fixed it, and I'm on to a more exotic error: 401 Client Error: Unauthorized for url: https://scholar.social/@abebab ; Request not signed
#
IWDiscordGateway <kongaloosh> running a curl post from "source=https://kongaloosh.com/followers"
#
IWDiscordGateway <kongaloosh> I think it doesn't check u-follow-of. Thanks for reminding me of the tool
#
IWDiscordGateway <kongaloosh> hmm, I got rid of abebab from my follow list and got a 202, sooooooo 🤷
#
[timothy_chambe] This is how the Dan Winer Activitypub ask seems to be going: https://twitter.com/tgruben/status/1589473026760134656
#
@tgruben @davewiner This is pretty well done. https://blog.joinmastodon.org/2018/06/how-to-implement-a-basic-activitypub-server/ (twitter.com/_/status/1589473026760134656)
#
@Corlaez ↩️ well some just bet on better interopability among simple personal websites (as in your own domain): microformats 2, webmentions, etc (twitter.com/_/status/1589479023482408961)
CrowderSoup and CrowderSoup4 joined the channel
#
@ellyloel ↩️ Oh! and it's hosted on @Netlify and it's using http://Webmention.io and http://Brid.gy for webmentions. (twitter.com/_/status/1589513605057482752)
#
@ellyloel ↩️ It's built with @eleven_ty using many awesome community made plugins, such as: @slinkitydotdev: @BHolmesDev Unfurl: @SaraSoueidan and @DavidDarnes Webmentions: @CodeFoodPixels Nesting ToC: @jshurmer Favicon Generation: @NJAldwin JSON Feed: @genehack Backlinks: @binyamingreen (twitter.com/_/status/1589513575860953089)
#
@ellyloel ↩️ Oh! and it's hosted on @Netlify and it's using http://Webmention.io and http://Brid.gy for webmentions. (twitter.com/_/status/1589513605057482752)
mro, petermolnar, CrowderSoup9, [Jamie_Tanna], barnaby, jbove and CrowderSoup4 joined the channel
#
@kirilind ↩️ you won't be able to keep track of all of them, I'm afraid - but you can do mutual linking between your website and your social profiles (which also turns some of them into #indieauth mechanisms, another #fediverse feat) to verify the "real you" in cyberspace (twitter.com/_/status/1589605340588683265)
[campegg], [campegg]1, [Dinis_Correia] and geoffo joined the channel
#
[snarfed] [Jamie_Tanna] https://github.com/snarfed/bridgy-fed/issues/267, just fyi
#
Loqi [snarfed] #267 Mastodon is rejecting our signatures for www.jvt.me
#
Zagura[m] Is is considered good practice to have an h-card in a sub-page, like example.com/about, but put a link in the main page (example.com) with the h-card class?
#
Zagura[m] I see an example in the official wiki, but im not sure if that means that the parsers will load the content of the link & extract the h-card from there, or just see an h-card with a single url
#
barnaby Zagura[m]: what example did you find?
#
Zagura[m] `<a class="h-card" href="https://tantek.com/">Tantek Çelik</a>, `
#
Zagura[m] here: https://microformats.org/wiki/h-card
#
Loqi Tantek Çelik
#
Loqi Tantek Çelik
#
[snarfed] sounds like https://indieweb.org/h-card#Issues
#
barnaby there isn’t currently any method specified for parsers to follow links from a light-weight h-card to a “full” h-card on a separate page
mro joined the channel
#
barnaby but in practise, most implementations which consume h-card are using it only for name, photo and URL, so a minimal h-card with those three properties will cover most existing usage
#
Zagura[m] damn, i was rly excited to shove all of my information into 1 h-card (not /s)
#
rubenwardy The wiki is misleading in this regard, I tried to do the same thing a few days ago
#
[dmitshur] So I'm learning a bit about ActivityPub and Mastodon, and want to understand more about "favorites" or "likes". Is it generally possible to look up who specifically favorited a given publicly-accessible toot/note/article, or does this vary based on some factors? I found a brief description of mastodon likes at https://blog.djnavarro.net/posts/2022-11-03_what-i-know-about-mastodon/#mastodon-favourites-vs-twitter-likes, and mastodon
#
[dmitshur] toots that seem to behave very differently: one where I can append "/favourites" to the URL and see a list of people (e.g. https://mastodon.social/@crawshaw/1567621/favourites), and another where that's 404 (e.g. https://inuh.net/@crawshaw/109302902834793494/favourites).
#
[snarfed] also https://microformats.org/wiki/representative-hcard , https://indieweb.org/representative_h-card
#
Zagura[m] anyway, thanks, i think ill keep the h-card stuff in the main page and use /about for more specific text stuff
#
barnaby oh good call [snarfed] I had forgotten about that 😅
#
Zagura[m] [dmitshur]: those are implementation differences: both instances might be using ActivityPub to communicate, but the ui & pages (including urls) could still be different
#
Zagura[m] only the special urls are the same afaik, like example.com/users/user_name
#
[dmitshur] Hmm. So if I understand it right: individuals own their "like" activity (and host that somewhere on their instance) and notify the server hosting a "toot"/etc.Then it's up to the server whether it exposes a detailed list of those who've favorited a given toot, or just a count. If the server doesn't expose detailed information, it's not really viable for an outsider to look it up.
#
barnaby [dmitshur]: yep, that’s correct
#
[dmitshur] Thanks.
#
barnaby the same applies for likes/favourites marked up with u-like-of and notified via webmention
jacky joined the channel
#
jacky what is birdwatch
#
Loqi It looks like we don't have a page for "birdwatch" yet. Would you like to create it? (Or just say "birdwatch is ____", a sentence describing the term)
#
jacky o.o
#
jacky looks like https://github.com/twitter/birdwatch holds some of the logic they use there for the birdwatch system (I've been using it)
#
jacky found an interesting note they had about 'diversity' https://twitter.github.io/birdwatch/diversity-of-perspectives/
#
barnaby what is birdwatching?
#
Loqi It looks like we don't have a page for "birdwatching" yet. Would you like to create it? (Or just say "birdwatching is ____", a sentence describing the term)
#
jacky lol close
#
barnaby everyone interested in the partial vs complete homepage h-card discovery issue, what do you think about my suggestion to use rel-about from the partial homepage h-card to an expanded about page h-card here https://indieweb.org/h-card#partial_vs_complete
#
barnaby seems like a no-brainer to me TBH
#
barnaby well, except for it being in the rel value table despite not appearing to be specified anywhere…
#
Zagura[m] i like it
#
barnaby capjamesg: did you mention using some sort of large-scale web crawl a while back for looking at mf2 usage? what was it again? I’d be interested to search it for rel-about usage if possible
#
[snarfed] https://httparchive.org/ ?
#
barnaby I think I remember seeing a different one, but that could be interesting too
#
barnaby ah it was commoncrawl.org
#
[snarfed] but beware https://indiemap.org/docs.html#common-crawl
#
barnaby huh did the rev attribute get dropped from HTML? it’s not in https://html.spec.whatwg.org/multipage/links.html
#
sebbu theorically we could just use https://query.wikidata.org/
#
barnaby ah good to know [snarfed], thanks
#
sebbu (in practice, relations aren't standardized/normalized, so it's complicated to use)
#
barnaby sebbu: how does using query.wikidata.org work for looking up existing link rel usage? I don’t see an example for anything similar to that
#
sebbu i once wanted to list the capitals of all states of a country
#
sebbu and it failed
#
sebbu (the links aren't always in the same direction)
#
sebbu (or with the same name)
#
sebbu beside, it also holds true for smaller administrative divisions, like districts and such
#
sebbu state, county, township, ward, etc... ( the right part in https://en.wikipedia.org/wiki/List_of_states_and_territories_of_the_United_States )
#
barnaby hmm I would look through the common crawl stuff myself but I don’t have 100TB of space available
#
sebbu "This article is part of a series on
#
sebbu Political divisions of
#
sebbu the United States"
#
sebbu you can query online
#
sebbu that's what i did
#
sebbu but i would have to list like 5 or 6 alternative for each relation, because it's not normalized at all
#
sebbu and i had several levels to go through
#
sebbu i quickly dropped the idea
#
sebbu it's almost faster to do it all manually
#
jonnybarnes snarfed I’ve posted in github, cross-posting works now
#
jonnybarnes super weird though, my database had token-a with all scope, and token-b with only read scope, clicking get token was giving me token-b, now gives me token-a, Mastodon didn’t need to regenerate a new token, just re-used token-a
#
[tantek]1 barnaby, rev was dropped ages ago. Check microformats wiki pretty sure
#
jonnybarnes oh lovely, github is being slow
#
barnaby [tantek]1: good to know, it was always a bit confusing. would be interested in your opinions on the rel-about stuff I linked earlier, I gotta go rehearse now though
#
[jgarber] barnaby: Somewhere either on the microformats wiki or the indieweb wiki there’s content about `rel=“me”` links to pages on the same domain being a lightweight flag for extended user data.
#
[jgarber] Trying to remember whether that’s part of the representative h-card spec, RelMeAuth, both… neither… Been looking at too many wiki pages too late in the evening to recall.
#
[jgarber] Ahh, some of that is noted here: https://microformats.org/wiki/RelMeAuth
#
Loqi Tantek Çelik
#
[jgarber] …though some of the examples are but long in the tooth or no longer accurate.
#
[snarfed] jonnybarnes good news, thank you!
#
[snarfed] and yeah that's interesting Mastodon token behavior. I guess I get it, seems fine
#
[tantek]1 That Anil dash post is spot on about no one fleeing Twitter to go to web3 or anything blockchain. You know what else people are not fleeing to? “Solid”
#
[snarfed] congrats btw jonnybarnes, that's the first successfully completed Bridgy Micropub post!
#
Saphire barnaby: > light-weight h-card with an href to a "full" h-card
#
Saphire JSON-LD flashbacks
#
jonnybarnes oh wow, history maker :P
#
Saphire Hm?
#
Saphire Oh heck, did not notice the discussion o:
gRegor and [jeffpaul] joined the channel
#
@schnarfed ↩️ This is possible today! https://brid.gy/ Built in support would be much better though. (twitter.com/_/status/1589681641303441408)
#
[Jamie_Tanna] Snarfed++ for super speedy https://github.com/snarfed/bridgy-fed/issues/267 fix
#
Loqi Snarfed has 29 karma in this channel over the last year (55 in all channels)
#
Loqi [snarfed] #267 Mastodon is rejecting our signatures for www.jvt.me
#
[snarfed] thanks! btw [aaronpk] this ^ was the sig verif failure I mentioned that was happening with both @-@ and URL key ids
#
[snarfed] unrelated to @-@ vs URL, and fixed now
#
@acegiak ↩️ Implementing activitypub/activitystreams gets you #fediverse integration. Implementing mf2 and webmentions gets you #indieweb integration. Easy. (twitter.com/_/status/1589684207265099776)
#
@megarush1024 ↩️ Fully support both the Activitypub and Bluesky protocols. Also, while we're at it, support open standards like webmention and micropub. (twitter.com/_/status/1589684306242375686)
#
@tchambers ↩️ Learn and integrate lessons and the open source bits from efforts like Bridgely https://brid.gy/about#publish and @microdotblog —- https://www.manton.org/2018/11/07/microblog-mastodon.html and http://Write.as platforms… https://www.manton.org/2018/11/07/microblog-mastodon.html (twitter.com/_/status/1589685353648033792)
[Sam_Butler] joined the channel
#
[Sam_Butler] I'm interested in people's opinions on htmx, and also aral's related work with domain(https://codeberg.org/domain/app), kitten(https://codeberg.org/kitten/app), nodekit(https://codeberg.org/nodekit/app), generally incorporating htmx and hyperscript. I saw a bit of recent discussion about aral's work in the meta channel and I'm curious to learn more. To me, these projects (and generally the energy from htmx/hyperscript) seems
#
[Sam_Butler] really exciting. Making things simpler for technical folks across a wide-range of applications, and making things more fun and accessible and approachable for everyday people.
mro, jacky, jordemort, h4kor[m] and [jeffpaul] joined the channel
#
h4kor[m] Finally done implementing IndieAuth 🥳
#
barnaby back
#
barnaby congrats h4kor[m]!
#
barnaby [jgarber]: re same-domain rel-me links, yep that’s true, and could work for multiple same-domain url properties on an h-card, too
#
[jgarber] An example that comes to mind is on aaronpk’s homepage: https://rel-me.cc/search?url=https://aaronparecki.com
#
Loqi rel-me.cc
#
barnaby there’s a link to aaronparecki.com/aaronpk, but that doesn’t really contain an expanded h-card, all the h-card data is the same as far as I can see
#
[jgarber] `/aaronpk` serves as his ActivityPub subscribe page, so there’s maybe not extended h-card info there, but a parser could see what could be retrieved.
#
[jgarber] Right, so it’s not a perfect example, but it’s the first that came to mind based on what I remember seeing from different people’s homepages.
#
[jgarber] I should probably stop using Aaron’s homepage as a test URL for all my various and sundry projects. :melting_face:
#
Loqi yea
#
barnaby when I was working on some XRay tests it was fun to see my domain being used as test data in there ;)
#
barnaby [jgarber]: thanks, I noted the same-domain rel-me possibiltiy with the relmeauth citations in the issues section
#
barnaby hopefully there’s a bit more there to work with now
#
[jgarber] barnaby++ Nice work, thanks!
#
Loqi barnaby has 26 karma in this channel over the last year (43 in all channels)
#
barnaby this has come up enough that it seems worth paying some attention to
#
@slyduda ↩️ also webmention supported crossposting via bridgy https://twitter.com/schnarfed/status/1589681641303441408?s=46&t=AfWgxqZOhmCXbXh3k-4nlg (twitter.com/_/status/1589717253918167040)
#
[jgarber] barnaby: I remember now I may be sitting on a dataset that could find some same-domain rel-me examples: https://github.com/jgarber623/indieweb-wiki-leaderboard/tree/main/data/users
#
[snarfed] and `pages.rels` in IndieMap: https://indiemap.org/docs.html#schema-pages
#
[snarfed] eg here's all rel values w/page count: https://console.cloud.google.com/bigquery?sq=464705913036:be3315d7205d4923a49721ca0af3b189
#
[snarfed] 150 About, 95 about, 3 About? , 3 about. , 3 about… 😆 . (out of 5.8M total)
#
barnaby hmm the link doesn’t seem to work for me
#
barnaby but thanks anywa [snarfed]
#
barnaby *anyway
#
[snarfed] hmm! you're logged into Google?
#
barnaby yep
#
[snarfed] interesting, ok
#
barnaby oh now it’s showing me a new screen with the option to “enable the API”, I’ll try that
#
barnaby now I get an editor with the query, but an error message saying “api-project-962594101220:indiemap was not found in location US ”
#
barnaby but I’ll take your word for it ;)
#
[snarfed] barnaby sorry! try again? might work now
#
[snarfed] not a big deal thoough
#
barnaby oh yeah that works! thanks [snarfed]++
#
Loqi [snarfed] has 30 karma in this channel over the last year (56 in all channels)
#
[snarfed] https://console.cloud.google.com/bigquery?sq=464705913036:8db3dcd2708e4b0c8cc2926beab2b0cd is same-domain rel-me links. 624 domains have them, out of 2300
#
[snarfed] https://snarfed.org/indiemap_rel_values_by_page_count.csv , https://snarfed.org/indiemap_same_domain_rel_me_links.csv
[Zeina] and t0nic joined the channel
#
[jgarber] This is _great_! snarfed++
#
Loqi snarfed has 31 karma in this channel over the last year (57 in all channels)
#
barnaby yes excellent data, thanks a lot!
#
[snarfed] welcome!
#
barnaby [snarfed]: is there a convenient way to list domains which have rel-me links on the same domain which are a different page to the one being linked? that would get closer to identifying cases of the same-domain rel-me expanded-h-card use case
#
barnaby I poked around with the other query and couldn’t get anything working, and I get the api project location error in the new one you linked
#
[snarfed] yeah that's in the query and the CSV, the same_url column
#
[snarfed] does https://console.cloud.google.com/bigquery?sq=464705913036:8db3dcd2708e4b0c8cc2926beab2b0cd work?
#
barnaby ah okay, thanks. I’m a little curious what adactio’s 171000 rel-me links are
#
[jgarber] => https://rel-me.cc/search?url=https://adactio.com
#
Loqi rel-me.cc
#
barnaby hmm so there aren’t actually any same-domain different-URL rel-mes on adactio.com? maybe there were at the time of the crawl?
#
[snarfed] the queried data is from crawl time, June 2017
[jeffpaul]1 joined the channel
#
barnaby and it looks like almost all of aaronpk’s pages have some rel=me links on, which look like they aren’t supposed to be there? e.g. https://rel-me.cc/search?url=https%3A%2F%2Faaronparecki.com%2F2022%2F11%2F06%2F51%2Ffitbit
#
Loqi rel-me.cc
#
[jgarber] Ahh, those are `<link rel="me" href="…">` on Aaron’s site.
#
[jgarber] …within `<body>`. 🤔
mro joined the channel
#
@theAdhocracy ↩️ Somewhere in the current experiments with Social Readers and webmention-native services like http://Micro.blog I feel like there is a really solid app, but nothing has quite clicked yet (twitter.com/_/status/1589737674352791557)
#
barnaby so I’m having trouble figuring out what if anything to conclude about same-domain different-URL rel-me from that data
t0nic joined the channel
#
barnaby [jgarber]++ for rel-me.cc, very handy tool. I assume it parses Link headers as well as markup?
#
Loqi [jgarber] has 4 karma in this channel over the last year (9 in all channels)
jacky joined the channel
#
barnaby so I used rel-me.cc to look at the top 20 domains in that list, and the closest thing I found to a same-domain separate-page h-card was https://rel-me.cc/search?url=http://boffosocko.com
#
Loqi rel-me.cc
#
barnaby so while I suspect it’s a more common pattern than rel-about, it’s hard to tell anything from that query specifically
#
[snarfed] ah that query looked at all pages, not just home pages
#
[jgarber] Aww, thanks. barnaby: I _think_ it’s only parsing markup. I’d have to dig back through the source.
#
barnaby [snarfed]: I know, but from the results I was only able to look for current homepage data to confirm what it’s reporting
#
[jgarber] Ahh, nope, just markup: https://github.com/jgarber623/rel-me.cc/blob/main/rel_me_app.rb#L90
#
barnaby and in theory rel-me is only useful on homepages anyway? e.g it looks like the sites with tens of thousands of results are mostly due to markup errors
#
[jgarber] …though it easily _could_ parse Link headers.
#
barnaby but it’s difficult to confirm that, at least with my (lack of) bigquery knowledge
#
barnaby [jgarber]: I can raise an issue for Link parsing if you want
#
gRegor Ooh, nice to see the partial v full h-card discussion. Haven't thought about that in a while.
#
[jgarber] Look what idiot wrote a gem for that, too: https://rubygems.org/gems/link-header-parser
#
[jgarber] barnaby: If you don’t mind, an issue on the repo would be great!
#
barnaby gRegor: a couple of people popped up in the last few days and were having issues with it, and were understandably frustrated about the inconsistent status of information about it
#
gRegor Trouble with it for authorship or something?
#
barnaby done!
#
barnaby gRegor: yep, both unclear how to do it, whether it’s supported, and if so how to test it
#
barnaby authorship, that is
#
gRegor Hm
#
barnaby so I tried to look into other alternatives and gather some data to inform discussion
#
barnaby personally I like rel=about for this as it’s explicit and clean, but I’m leaning towards same-domain rel-me as it’s likely more widely used
#
[snarfed] barnaby one example, https://huffduffer.com/snarfed, the "Older" link to https://huffduffer.com/snarfed?page=2 has rel=me
#
gRegor So proposing extending https://indieweb.org/authorship-spec#Algorithm to support rel-about?
#
barnaby I’m just wondering if “same domain” is always going to be a useful heuristic
#
barnaby [snarfed]: ahh, that would explain the vast quantities of huffduffer rel-me links! seems like a misuse of rel-me to me IMO, if it’s for pagination
#
[snarfed] another, singpolyma's posts all rel-me his home page, eg https://singpolyma.net/actionstream/will-miss-mostall-of-wlsg/
#
barnaby gRegor: I think the best method would be an extension to the representative h-card algorithm, to hint that if consumers find a representative h-card with same-domain url properties or rel-me values, they could fetch those to potentially find an expanded h-card
#
barnaby because the authorship algorithm will lead to something which the representative h-card algo can be done on, so we only need to update one algorithm that way
#
gRegor Ah, ok. That makes sense. Authorship was getting me confused.
#
barnaby [snarfed]: yep, arguably another misuse of rel-me, and the one same-domain one (https://singpolyma.net/contacts/) isn’t an expanded h-card but a contact list
#
barnaby I’m specifically trying to find wild examples of same-domain rel-me links from a homepage to a page with an expanded h-card on
#
[snarfed] ah ok got it, I can try to query that in a bit
#
barnaby that’d be great, thanks a lot!
tbbrown joined the channel
#
[tantek]1 barnaby, cross-domain would violate CORS and such
#
[tantek]1 hence why rel-me requires a bidirectional link across domains for it to mean anything
#
barnaby yeah I’m not even thinking about cross-domain rel-me links here
#
[tantek]1 sounded like you were considering it re: about pages
#
barnaby nope, never crossed my mind
#
[tantek]1 I don't see what user-problem rel-about is solving
#
[tantek]1 that isn't already solved by existing rels
#
[tantek]1 barnaby, got it
#
[tantek]1 not just same-domain, but rel-me one-way links only work when they're deeper into the same path
#
barnaby my thinking in rel-about vs same-domain rel-me is that rel-about is a clear explicit indication that “the single linked page provides more information about me”, which is a *very* common use-case, whereas using rel-me for that mixes several use cases and adds additional necessary heuristics, e.g. same-domain or URL prefixes
#
barnaby and yeah, specifying to look for expanded h-cards on rel-me/u-url links which are prefixed by the h-card URL could help avoid this causing issues in cases where there are multiple profiles per domain
#
barnaby “how can I have a minimal h-card on my homepage but an expanded one elsewhere” is a question which comes up quite often e.g. twice within the past few days
#
barnaby so that’s the user need, and there isn’t yet a clear answer for it
#
barnaby that’s why I was trying to look into existing usage of rel-about, to see if there’s a clear pattern of people using it to mark up this publishing use-case already (unclear, but it is definitely in use one way or another https://indieweb.org/h-card#partial_vs_complete)
#
barnaby and am now trying to do the same for same-domain rel-me
#
barnaby so if same-domain rel-me works for this use case, then I suggest we draft and test an extension to representative h-card for consumers to follow if they’re not satisfied with the representative h-card found (e.g. a partial homepage h-card) and want to find an expanded one
#
[snarfed] barnaby++, more of us had this use case (I still do) and thought through it a bit a long time ago in https://indieweb.org/h-card#Issues
#
Loqi barnaby has 27 karma in this channel over the last year (44 in all channels)
#
[snarfed] (I have no idea of the right mf2 plumbing though)
#
barnaby [snarfed]: ah yeah I see you have an expanded h-card on https://snarfed.org/about
#
[snarfed] yup
#
barnaby and what the right mf2 plumbing would be still remains a question, yeah. at the moment I’d say having https://snarfed.org/about as a secondary u-url, and additionally rel-me when linked from your homepage would make sense
#
barnaby that’s all that would be required for the same-domain/prefixed rel-me heuristic to find it
#
[snarfed] ok, indie map has 29 of these. https://console.cloud.google.com/bigquery?sq=586366768654:047463741d5343c3ab218b152561b005 , https://snarfed.org/indiemap_rel_home_page_to_another_page.csv
#
[snarfed] definitely seeing many of those have rotted now
#
[tantek]1 barnaby, the same domain / path prefix thing has nothing to do with the rel value, it's about can you actually trust the assertion being made by the site about a relationship at all. regardless of "about" "me" "flurgle" or whatever
#
barnaby I never asserted that same domain has anything to do with rel value? I’m trying to assess arguments for how to mark up this common publishing pattern
#
barnaby thanks [snarfed], very interesting!
#
barnaby interesting to see so little usage of this. I wonder what the >200 usages of rel=about/rel=About in the same dataset were being used for
#
barnaby and regardless of how it’s marked up, finding expanded h-card information on a different domain would be trustable if it linked back, as with all rel-me usage. so the rel-about solution would actually allow publishing your expanded h-card on a separate domain, not that that appears to be useful (or something we want to encourage!)
#
barnaby yeah it looks like not a single one of thsoe 29 examples is actually a still-working expanded h-card
#
barnaby or even a still-working about-me page regardless of h-card markup
#
[snarfed] actually sorry I undercounted, the real number is more like 115 ish if you count subdomains like www, blog, etc
#
barnaby ah okay, that sounds promising wrt finding working examples
#
[snarfed] updated https://snarfed.org/indiemap_rel_home_page_to_another_page.csv
#
barnaby I wonder if it’d be worth just querying domainname.com/about and seeing how many show up in the dataset? that’s a very common URL to use for this use-case, and as we can see from snarfed.org not all of them are marked up with rel-me
#
barnaby thanks! reviewing now
#
[snarfed] sure! that's https://console.cloud.google.com/bigquery?sq=586366768654:515ee576e5f348688c809723f7078088
#
[snarfed] (skipping CSVs now if you're able to see these queries)
AramZS joined the channel
#
barnaby yep, the queries are all working for me now, thanks!
#
[snarfed] hopefully the SQL makes sense, feel free to poke at it yourself! more background in https://indiemap.org/docs.html#data-mining
#
barnaby yep, I think I’m slowly getting a sense of how it works. never worked with bigquery before and my attempts to tweak your queries to do what I wanted didn’t work out so far xD
#
[snarfed] ah sorry. basic SQL should work! some of the columns are repeated and nested there are little quirks to getting those, feel free to browse the existing project queries for examples
#
[snarfed] https://cloud.google.com/bigquery/docs/reference/standard-sql/introduction
#
aaronpk oops do i have extraneous rel me markup?
#
barnaby aaronpk: looks like it 😅 you’re far from alone, though