#dev 2021-03-22

2021-03-22 UTC
nertzy, [Jacob_Hall], [chrisaldrich] and [snarfed] joined the channel
#
[snarfed] [gRegorLove]
#
[snarfed] oops misfire
#
[snarfed] [gRegorLove] just fyi you’ve been fetching https://brid.gy/like/instagram/gregorlove/2533370160718255387_210105230/520 at about 1qpm for a while now, just a heads up in case it’s a bug
#
[snarfed] not a big deal!
treora, [tantek], [aimee] and Zenyattus joined the channel
#
@kisik21 ↩️ Use a lot of microformats and either CI magic or JavaScript for webmentions. That's how I did it I guess. (twitter.com/_/status/1373893528431038464)
[KevinMarks] joined the channel
#
@inautilo Adding Webmention support from scratch · What to do if you can’t use a CMS for Webmentions https://ilo.im/rqgm3 #guides (twitter.com/_/status/1373971684500324358)
IWSlackGateway and [snarfed] joined the channel
#
[snarfed] ugh JSON-LD. any of you who’ve worked with ActivityPub, curious to know if i’m missing anything here: https://github.com/snarfed/bridgy-fed/issues/76
#
aaronpk wtf when did that change
#
aaronpk I really don't understand why JSON-LD is necessary here... just use JSON
#
sknebel gonna say, quite possible it didn't "change", it's just that nobody cared about it being consistent JSON-LD
#
aaronpk Maybe? I feel like we would have heard more about it in the early days and seen it in the spec examples and I definitely don't remember seeing that URL before
#
[snarfed] ugh it’s an HTTP sig thing. goddamn AP authz. https://github.com/w3c/activitypub/issues/203#issuecomment-297553229
#
sknebel publickey isn't in the spec though?
#
sknebel (afaik)
#
aaronpk ah right
#
[snarfed] not the AP spec, no. the authz part of the AP spec basically just says “🤷 yolo!”
[tantek], dckc and [scojjac] joined the channel
#
[tantek] the right answer IMO is to drop @context everywhere because the claim of JSON-LD advocates is that you can have compatibility without it. any incompatbilities should be pushed back onto the JSON-LD processing model crowd
[Rose], [KevinMarks], baslow and sumner joined the channel
#
[KevinMarks] Ah, I thought that I could delegate micropub to known.kevinmarks.com from kevinmarks.com but they have different indieauth endpoints, and that gets rejected. I'll have to try something craftier to get the full social reader experience
#
aaronpk yeah the claim/justification for keeping JSON-LD in activitypub was that all complexity would be pushed onto people who care about it and others would be able to treat it as "just JSON"
#
[tantek] exactly ^
#
[tantek] that's the reason to hold firm to ignoring / not publishing @context
#
[KevinMarks] And they can amend the implied context to include whatever other namespaces that they want?
#
[tantek] if that's the sort of thing you even care about
#
[tantek] I'm just waiting for schema-org to declare a canonical ns for this stuff that then everyone RDFy that cares uses which of course will the obviate any need for it in the first place
[grantcodes] joined the channel
#
[snarfed] [tantek] thanks for the nudges. any chance you have a link for that compatibility claim for JSON-LD advocates? i can try pushing back with that
#
[tantek] snarfed, no specific link, just every time it came up discussions
#
@JeremiahLee This is a harsh way of fixing a typo. But seriously, I need to figure out pingbacks and webmentions now. (twitter.com/_/status/1374031147055910918)
[chrisaldrich] joined the channel
#
sknebel I kinda guess nobody really considered the extra-contexts case and what exactly that means. I.e. from my outside perspective it read pretty much as "lets agree on a form that's valid JSON-LD and easy to handle without JSON-LD (i.e. dont require any weird resolving etc)", and perception of the latter focussed mostly on *consuming*, not "it's actually a bit more than just copying the same template around everything once you leave the
#
sknebel specified space and need extra contexts you need to add those"
#
sknebel (afaik it doesnt hurt to have unneeded ones, so it can still be a global template, but not just the one from the spec)
Guest78 joined the channel
#
aaronpk there's probably a reference to it in various meeting minutes
#
aaronpk https://www.w3.org/wiki/index.php?title=Special%3ASearch&profile=advanced&fulltext=Search&search=%22just+json%22&ns0=1&ns102=1&ns108=1&profile=advanced
shoesNsocks and [KevinMarks] joined the channel
#
[tantek] wow I'm getting a "503 Service Unavailable" in a completely unexpected part of my personal site UI and I'm really curious how that's possible in PHP (no iframes obv)
#
[tantek] (and I thought I'd coded this part of my site *very* defensively, so it's a pretty big surprise that a. that would happen, and b. that such an "error message" would actually make it through from the backend and leak into the frontend)
#
sknebel something times out or your host has issues
KartikPrabhu joined the channel
#
[tantek] in this case it is returning (rendering the page) very quickly so ...
mxd joined the channel
#
mxd how should i markup comments to a post which are created one the same page to receive/send webmentions for the post? With "h-entry" and "u-in-reply-to" and then send a webmention with the same url for target and source?
#
mxd Eh nvm😅
#
[tantek] mxd, like how to mark up a /comment post? or how to mark-up /comments shown under a post of yours?
#
mxd Second one. Perfect that's what I was looking for. Thank you
#
dansup_ I finally launched https://fedidb.org :)
#
dansup_ Need to move it to a bigger server though, didn't expect it to be posted to HN so soon
#
[tantek] ugh and now the 503 is no longer showing up
#
[tantek] transienterrors--
#
Loqi transienterrors has -1 karma over the last year
#
aaronpk "funded by NGI Zero" interesting
#
aaronpk dansup_: any chance a similar tool for webmention/micropub might be covered under that grant too? :)
[jeremyfelt] and [tw2113_Slack_] joined the channel
#
dansup_ aaronpk: Yeah probably! Pixelfed received two NGI grants, FediDB is part of the second grants milestones
#
[tantek] does anyone create top level paths as redirects for particular silos?
#
[tantek] e.g. I do this with github so I can share URLs like https://tantek.com/github/cassis
#
[tantek] and thus own-my-links as it were with my own domain, allowing myself the flexibility of changing where I actually host that stuff
#
[tantek] could be an straightforward way to pick URLs to host backfill stuff from silos too
[snarfed] joined the channel
#
[snarfed] tantek++
#
Loqi tantek has 18 karma in this channel over the last year (80 in all channels)
#
[tantek] asking for other examples, also have no idea what to name this "practice" in order to create a page about it
#
[tantek] maybe this is a broader question about site information architecture (IA) and top level paths as a whole?
#
[tantek] some of this may have been subsumed by the pages vs posts debates?
#
[tantek] what are pages
#
Loqi It looks like we don't have a page for "pages" yet. Would you like to create it? (Or just say "pages is ____", a sentence describing the term)
#
[tantek] ok then
#
[tantek] hmm I suppose that could mean multiple things
#
[tantek] going to make something up and see if it sticks / resonates with anyone
#
[tantek] pages is a place on your website where you list and link to individual (typically top level) [[page]] URLs on your own site, sometimes part of top level [[navigation]] or a list of links on a [[homepage]].
#
Loqi ok
#
mxd awsome idea. just implemented it. https://mxd.codes/github/DaTurboD. gonna replace all links to silos with it.
#
mxd tantek++
#
Loqi tantek has 19 karma in this channel over the last year (81 in all channels)
#
Loqi Max Dietrich
#
[tantek] mxd thanks! do you have a suggested name for this practice/idea? 🙂
#
mxd https://mxd.codes/github/DaTurboD
#
Loqi Max Dietrich
#
mxd guess you could call it somehow like wildcard-redirects
#
mxd or path-matching
#
[tantek] we do have /ownyourdata and maybe this could be a variant (subset?) of that with "ownyourlinks"
#
[tantek] kind of a meta question but it's devy enough that I'd like more dev-thinking thoughts
#
mxd sounds good to me 👍
[KevinMarks] and KartikPrabhu joined the channel
#
[chrisaldrich] I've seen people do a variation on that pattern before tantek. There were many in the days of Google+ who would either redirect their entire personal domain to Google+ or do it with an example.com/google+ link. I'll see if I can find examples. Don't remember anyone having a name for the broader pattern other than just a vanilla redirect.
#
[chrisaldrich] I almost want to say there was maybe a reason to do the /google+ pattern from Googles side, but don't recall it now. I want to say I may have done it at the time, but must have gotten rid of it when G+ died??
#
[KevinMarks] there was authorship where they would show your g+ pic in search results for a while
#
[KevinMarks] the original google profiles (wot i built) was pretty much a list of rel=me links
#
[chrisaldrich] Here's an historical example of the first pattern of redirecting one's entire site to G+. (I'm guessing it didn't work out well in the long run---since it turned into a starvation diet in the end.) https://hyp.is/6ZP58ItREeuz5q_fZnlY_Q/www.computerworld.com/article/2510464/elgan--what-i-lost-on-the-google--diet.html
#
sknebel I see it occasionally for project pages of open source projects. i.e. "<project>.<tld>" or "<project>.<tld>/code" being the URL shared and a redirect to the repo location
#
[tantek] redirecting entire domain seems like a bad idea
#
[tantek] this is only about top level pages
#
@dev_nikema Webmentions – let’s go! https://nikemaprophet.com/webmentions-lets-go/?utm_source=ReviveOldPost&utm_medium=social&utm_campaign=ReviveOldPost (twitter.com/_/status/1374116187341271041)
[KevinMarks], Seirdy and [scojjac] joined the channel