#dev 2022-11-26

2022-11-26 UTC
#
IWDiscordRelay
<j​acky#7226> The concept of TOFU is helpful in some contexts
geoffo joined the channel
#
Loqi
[mx alex tax1a - 2020 (2)] someone on Twitter asked why we think ActivityPub sucks from a protocol and implementation standpoint, and we're porting it over here with minor cleanups:several things — from a server perspective, the most popular implementation requires that you ...
#
aaronpk
oof "we both worked for twitter for 6 years and run our own **email server* for 21"
#
aaronpk
they missed the sharedInbox thing tho apparently, which is only marginally an improvement
#
barnaby
that’s the first instance of “mastodong” I’ve seen where I’m not entirely sure it’s not intentional
#
[manton]
[aaronpk] I only implemented sharedInbox when my servers start getting slammed the last couple of weeks, but unfortunately Mastodon seems to cache the user profile info for a long time… Most requests still aren’t using sharedInbox.
#
aaronpk
oh gosh
#
aaronpk
you can send out profile update activities if you want to force an update
#
aaronpk
tho now that i think about it, i'm not sure if that also updates the sharedinbox property. it definitely updates the name/photo/banner tho
#
GWG
Weird question...the length of a Syndication Target.
#
GWG
I'm adding Micropub Bridgy support alongside Webmention Bridgy Support. I'm afraid they'll get large
#
IWDiscordRelay
<c​orlaez#3273> Hi, it has been a while but I have been pondering something. mf2 allows parsing of a limited number of web relevant data... but is there something similar in spirit (as in attributes or CSS in html) that allows you to parse generic data (extract data from html). What are the most popular standards for that? I am sure there must be some
#
[manton]
[aaronpk] Oh thanks, I didn’t think about sending out an update. I should probably do that when someone changes their profile photo.
geoffo joined the channel
#
GWG
[manton]: Got a moment for me to pick your brain?
#
GWG
About Syndicating to Micro.blog
#
GWG
I'm looking at all my POSSE code.
#
GWG
The Micro.blog one creates a hidden custom feed that items can be added to in order to appear in micro.blog, just wondering there is anything I can do to improve it
neceve and gxt joined the channel
gxt and mro joined the channel
#
[KevinMarks]
Corlaez: depends what you're trying to do. Beautiful Soup will do a good job of letting you pull specific bits out of a page, but you are going to have to maintain a lot of mappings. There's granary.io which parses a lot of places into a common format, and there's https://github.com/postlight/parser which works on a lot of news sites. There's https://indieweb.org/XRay too.
#
[KevinMarks]
If you want previews rather than full content then parsing for Facebook metadata can work. Some libraries fall back on that if there's no mf2 support.
#
Loqi
[postlight] parser: 📜 Extract meaningful content from the chaos of a web page
barnaby and mro joined the channel
#
ash[m]
gRegor: I'm hoping I've fixed my markup for reposts so it's parsing correctly for you now?
#
ash[m]
oh hmm nope I'm looking at it now and it's still weird
#
barnaby
yep, looks like there are h-entries getting embedded in h-cites
#
barnaby
removing the redundant h-entry classname should end up with all the repost properties moving to the h-cite as expected
#
barnaby
(almost typed h-cute there lol)
#
barnaby
looking good to me now!
#
barnaby
hmm s/the test suite/monocle isn’t picking up you as being the reposter though https://monocle.p3k.io/preview?url=https://acegiak.net
#
aaronpk
what is authorship?
#
Loqi
authorship is how to indicate who the author is for a post, and an algorithm that determines the author of a post https://indieweb.org/authorship
#
aaronpk
there's no author property on the h-entry's there, i think that's why
#
barnaby
the notes you post have authorship info which gets picked up, but not the reposts
#
barnaby
putting an <a class="u-author h-card"> around the <img class="tiny-actor-icon"> on reposts might fix it?
#
aaronpk
i do see an author property on the non-reposts
#
aaronpk
(i'm just looking at the parsed mf2)
#
barnaby
the embedded h-cites have authorship, but not the repost itself
#
aaronpk
yep that should do it
#
barnaby
so monocle shows “someone reposted x” rather than “acegiak reposted x”
#
barnaby
looking good other than the hyphen in h-ref now ;)
#
barnaby
oh and you don’t need an explicit span.p-name there thanks to implied properties
#
ash[m]
I do cause otherwise monocle picks up the avatar alt text
#
ash[m]
think I've got it pretty good now. including marking up the attachments in the reposts!
#
barnaby
IMO if you don’t like how the alt text looks, it’s probably the alt text which needs improving
mro joined the channel
#
ash[m]
the alt-text on my avatar is "ash mcallan's avatar", which is why I then put the p-name separately?
#
ash[m]
it's not great alt text but if I just put "ash mcallan" as the alt text then I assume screen readers would be like "ash mcallan ash mcallan"
#
barnaby
currently screen readers would say “ash mcallan’s avatar ash mcallan”, is that much better?
#
ash[m]
right? I don't know
#
barnaby
afaik in contexts like this where the alt doesn’t describe the image or really add anything it’s better left out
#
ash[m]
yeah fair
#
ash[m]
I might remove the alt then
#
barnaby
and IMO that’s one of the hidden advantages of implied name-from-alt parsing: it makes the alt text visible to people who don’t use screen readers, and helps surface issues like this
#
barnaby
that’s what I’d do, yeah, with the caveat that I’m not an accessibility expert or screenreader user
#
barnaby
but afaik it’s the best practise in this situation
#
barnaby
I gotta ask Joschi about this at some point, it’s come up a few times
mro joined the channel
#
Loqi
[@ash@acegiak.net] Accessibility friends. If the alt text on an image isn't really adding much, ie describing non-interactive UI elements is it better to just remove it? Ie if I've got an avatar image next to a person's name, is it better to mark that up with an alt te...
#
barnaby
ash[m]: btw once you have one explicit property classname under a root h-* classname, implied parsing no longer happens. So adding the p-name without also adding explicit u-url and u-photo (and usually an additional container element if the h-card was on an <a>) means you lose them
#
ash[m]
Ahhh that explains that
#
ash[m]
Cool
#
barnaby
implied parsing for name, url and photo used to take place separately for each property if it wasn’t explicitly defined, but that resulted in way too many false positives to be useful
#
barnaby
might need to add a FAQ for this if we don’t have one already, it comes up fairly often and is understandably confusing
#
ash[m]
ok I've removed both the explicit pname and the alt-text
#
ash[m]
yeah i read the h-card page on the wiki but only in adhd mode so i don't know if it's explained in there but in a hidden way?
#
barnaby
there were also outdated examples on the wiki up until recently
#
barnaby
which didn’t help. I think I corrected them all
#
barnaby
markup is looking good to me now! ash[m]++
#
Loqi
ash[m] has 1 karma in this channel over the last year (2 in all channels)
#
ash[m]
Now I just have to make sure I don't break it next time I pull from main 😝
mro and oenone joined the channel
#
ash[m]
Pushed a patch at least. Ball is in tsileo's court now
mro and MerlinStar joined the channel
#
@endgameviable
↩️ Incidentally the more I work on this, the more I find other work on the Internet that already does blog-to-activitypub interop e.g. I just learned about https://fed.brid.gy/
(twitter.com/_/status/1596546391127597061)
mro, chenghiz_, gxt, geoffo and mro_ joined the channel
#
Loqi
[Christopher Schmidt] Fun fact: sharing this link on Mastodon caused my server to serve 112,772,802 bytes of data, in 430 requests, over the 60 seconds after I posted it (>7 r/s). Not because humans wanted them, but because of the LinkFetchWorker, which kicks off 1-60 sec...
#
[snarfed]
they added a random delay btw 1 and 60s to try to help
#
Loqi
[snarfed] i noticed this recently too. >1k requests in <45s, >25qps. not a disaster, my site handled it fine, but still, noticeable. small thread on it here: https://mastodon.technology/@snarfed/100119606571241751 , cc @ashfurrow @neekz0r. <img width="530" ...
#
[snarfed]
unrelated, whoa, I just noticed that huffduffer regularly crawls rel-me links and shows them on your profile. eg I only gave it my web site, but it found everything else: https://huffduffer.com/snarfed
#
[snarfed]
adactio++
#
Loqi
adactio has 1 karma in this channel over the last year (2 in all channels)
#
[snarfed]
oh man also, G+ 😂
#
barnaby
mastodon is the new slashdot
[jacky], geoffo and sp1ff joined the channel
#
[tantek]
what is thundering herd
#
Loqi
It looks like we don't have a page for "thundering herd" yet. Would you like to create it? (Or just say "thundering herd is ____", a sentence describing the term)
#
[tantek]
^ paging [KevinMarks] for the original feed file AKA "air quotes" RSS problem and adding Mastodon, perhaps fittingly named, to the thundering herd problem
#
[tantek]
snarfed, pretty sure AS2 has both follow and unfollow activities and they are different
#
[tantek]
might be useful as a personal record to keep track of when you followed/unfollowed folks, perhaps multiple times, e.g. as private posts except for the most recent (un)follow of another site
#
barnaby
IIRC it’s a follow and an “Undo”
gxt joined the channel
#
[tantek]
is it really? 🤦
#
barnaby
no weirder than “Announce” for repost
#
[tantek]
should just make an "un" modifier for any post type 😛
#
[tantek]
yeah "Announce" was weird and I think we even pointed it out in the WG, and no one wanted to do the work to update it to what people actually use (repost)
#
[tantek]
every time I see that use of "Announce" I hear/think 🎺 🎺 🎺
mro and [snarfed] joined the channel
#
[snarfed]
yeah it's an Undo. I was asking specifically for IndieWeb unfollows
#
[snarfed]
following the pattern of Bridgy [Fed] following existing, ideally common mf2, not pioneering it
gxt joined the channel
#
sknebel
https://simonwillison.net/2022/Nov/26/productivity/ this by simonw is quite interesting about how he makes it easier for him to go back to old code. definitively a problem I recognize e.g. from my site code :D