#dev 2021-03-31

2021-03-31 UTC
[scojjac] joined the channel
#
[tantek]
vilhalmer (from #indieweb) re: h-, see https://microformats.org/wiki/microformats2-prefixes (which I realize is long) and in particular "The 'h-' prefix is based on the existing microformats naming pattern of starting with 'h'."
#
vilhalmer
I found that, was wondering if there was any more lore :)
#
[tantek]
there's much more, much older, let me dig a little
#
vilhalmer
my gues would be vCard as html = hCard and they grew from there?
#
[tantek]
here you go, I knew I wrote this down: https://microformats.org/wiki/hcard-history#2004
#
[tantek]
short version: hCard = vCard in HTML, hCalendar = iCalendar in HTML.
#
[tantek]
so the "h" came from "in *H*TML"
#
vilhalmer
nice, thanks
#
aaronpk
I always thought it was an upside down µ but that makes more sense
#
[tantek]
yeah that was a backronym of sorts, and I think KevinMarks found it appealing so let it propagate 😛
#
KartikPrabhu
maybe "o-" for object would be better but oh well too late now!
#
[tantek]
objects are overrated KartikPrabhu
#
KartikPrabhu
then "o-" for overrated :P
[jgmac1106], Seirdy and KartikPrabhu joined the channel
#
@GallantIdeas
I've updated http://gallant.dev, now with comments, webmentions, and other #indieweb goodies. https://gallant.dev/posts/a-blog-reborn/ Expect continued irregular posting and unnecessary verbiage.
(twitter.com/_/status/1377088543504224259)
dckc, KartikPrabhu and [KevinMarks] joined the channel
#
[tantek]
KevinMarks I miss the rate of optimistic innovation & building we had in 2004-2006. And I wish we had the understanding (we have now) to deal with the nasty trolls back then
#
[tantek]
We gave email way too much of a chance
[Kai_Vong] joined the channel
#
[Kai_Vong]
there is a book: Microformats Made Simple, Emily P. Lewis and a tech review by [tantek] Çelik
#
[Kai_Vong]
↩️ there is a book: Microformats Made Simple, Emily P. Lewis and a tech review by [tantek] Çelik
deathrow1 joined the channel
#
@tomlarkworthy
After I brief detour for #WomansHistoryMonth and helping a friend with their startup I am BACK TO INDIEAUTH/RelMeAuth
(twitter.com/_/status/1377152588311715844)
dhanesh joined the channel
#
sebsel
jacky: It's not out of the blue that they jump on the Github ship. A group of people already wanted this for a long time and this gave them the excuse
#
sebsel
or reason.
#
sebsel
(this is the impression I have, I'm not that involved*)
[scojjac], [Rose], [barryf], [KevinMarks], [tantek], marinin[t] and [manton] joined the channel
#
[manton]
I like the proposed “beyond Webmention.io” pop-up session. I’m sure the focus will mostly be on self-hosting, but I’m also curious if there can be more services that are compatible with Webmention.io. For example, I have a feature in Micro.blog that will return the same JF2 as Webmention.io. I’d like to do more like that.
#
jacky
[manton]: my feelings are going to be around building services that use Webmention tbh
#
jacky
like hopefully someone being the vanguard on how we do Vouch
#
aaronpk
oh nice i missed that
#
jacky
for those reading and curious to what that session is: https://indieweb.org/2021/Pop-ups/Sessions#Webmentions_Beyond_Webmention.io
vilhalmer joined the channel
#
sebsel
it does not yet have a date, it seems? it looks interesting
#
jacky
not yet!
ShadowKyogre joined the channel
#
[manton]
[jacky] That makes sense. It just made me think about how to kind of spread the load from Webmention.io to other sites, or make it easier to move between services.
#
jacky
oh that's a good point too (reminds me about having a shared/common interface for fetching Webmentions - hopefully can be used as said export mechanism)
#
aaronpk
yeah i'd love to see us standardize on a front-end api for retrieving webmentions so people can make libraries and stuff that works with things other than webmention.io
#
aaronpk
and then someone please make a small Go project people can spin up to be their webmention endpoint :)
#
aaronpk
wow i really need some better logic around handling timeouts and stuff for my activtypub delivery. it's kind of a mess right now and depending on the failure mode, one person's server might hang the rest of my delivery for like 30 seconds
#
[manton]
Yeah. And I guess one question is whether any of that should be part of the Webmention spec itself, or just a separate common interface that a lot of people use.
#
sebsel
in a docker container so you don't have to? ;)
#
aaronpk
sebsel: if you want to use docker yes, but i refuse to :D
#
aaronpk
[manton]: anything like this is likely to be in a separate spec, but that's not as important as getting it working in the first place :)
#
sebsel
aaronpk: I'm currently not on docker either, but I know you get the request sometimes
#
sebsel
yea, the webmention spec currently doesn't say much about likes vs mentions vs replies and stuff, that's all Microformats.
#
sebsel
it seems logical that the common interface would say stuff about likes, because that's the number one feature: twitter likes
#
sebsel
but +1 for making it work first.
[grantcodes] joined the channel
#
[grantcodes]
I love docker for running random tech stuff. I really do not want to learn how to install dependencies, run builds, bind to ports etc etc in whatever language the project is written in.
[tantek], [Rose] and [schmarty] joined the channel
#
[schmarty]
still thinks Vouch is plumbing in search of a use case but excited to see some discussion about person-centric ways to mitigate the issues vouch purports to solve.
#
aaronpk
same tbh
#
[tantek]
Vouch was a first attempt at solving the use-case of trying to allow for new/interesting responses without being burdened with moderating everything
#
[tantek]
problem is we never got far enough with actual usable/interoperable implementations to get a feel for how well it actually solved that use-case, so it's kinda been stuck at the level of "interesting plumbing, worth considering?"
#
[tantek]
aaronpk, pretty sure you were there for that session in, I think 2013, at IWC MIT / Cambridge where we discussed the person-centric use-cases first before figuring out *a* protocol that *might* solve them
#
aaronpk
sounds familiar
#
[tantek]
[schmarty] I think you're right that we need to level up back to discussing the actual use-cases
#
aaronpk
problem is that technical solution stuck around in the general discourse more than the discussion did
#
[schmarty]
I bring it up whenever Vouch comes up 😂
#
aaronpk
it does sound like a different discussion than what is currently proposed for that webmention popup discussion tho
#
[tantek]
this is why I like to half-jokingly refer to the problem area instead as the "House Party Protocol"
#
[tantek]
it's much easier to relate to the (pre-pandemic at least 😞 ) problem of throwing a house party where you do want it to be semi-open to *some* folks you don't directly know
#
jacky
[schmarty]: do you have an idea for something else?
#
jacky
I'm open to ideas
#
jacky
I couldn't think of any immediately and vouch felt familiar conceptual
#
[tantek]
!tell benwerd might be good timing to revive your indieweb side-project fatberg.org, just saying: https://www.bloomberg.com/news/articles/2021-03-26/pandemic-wipes-create-sewer-clogging-fatbergs
#
Loqi
Ok, I'll tell them that when I see them next
#
jacky
like in a physical social env kind of way (5 people in a room, three pairs of them know each other so the ones out are then introduced by the common bridges)
#
GWG
I have a few ideas for webmentions
#
[tantek]
(there that'll be a nice easter egg for when he returns 😈 )
#
[tantek]
I like that comparison jacky
#
[tantek]
Twitter has a feature where you can set a tweet to only allow replies from people you follow. What would be even more interesting is if you could allow replies from people that *they* follow as well
#
[schmarty]
jacky: I think there are some good ideas to explore around "social graph"-like relationships for sure! I think Vouch in particular is too much about protocol and not enough about user experience.
#
[tantek]
I might actually set that as a default ("people you follow" is too narrow/stifling for me)
#
jacky
[schmarty]: good point
#
jacky
I do think it's meant to be a bit invisible to people though but I might stew on that a bit more (on how it'd look from people first)
#
[tantek]
[schmarty] it's likely we didn't capture the use-case / desired UX well enough at the initial brainstorming in 2013, and since then as aaronpk pointed out, the discussions have been so protocol-centric that they lose focus of the use-case 😕
#
[schmarty]
Vouch is a bit contradictory on how transparent it is depending on which part you are reading 😅
#
[tantek]
jacky, yes it's meant to be invisible as a protocol, but almost certainly indirectly visible in a simpler form in the UX
#
[schmarty]
Sometimes it assumes it is part of a healthy moderation toolset
#
[schmarty]
Sometimes it assumes a person is in the loop
#
[tantek]
like a feature for "allow responses only from people I follow or have linked to and people they follow" (not a great description, but that's kind of an intent)
#
[schmarty]
Sometimes it assumes a webmention sender (or endpoint) knows about every link to and from a particular person's posts
#
[tantek]
[schmarty] it was meant to be seamless most of the time, a way to reduce moderation tax, and not require a person in the loop
#
[tantek]
not "every" but "enough"
#
[tantek]
it's supposed to be put (some) labor on the side of the sender, and enable (more) automation on the side of the receiver
#
jacky
quite a bit to think on
#
[schmarty]
I think a recent time talking about Vouch we ended up discussing some big changes to approach, like dropping the assumption that webmentions are open by default
#
jacky
oh I wanna find that
#
[schmarty]
OH also a huge one for Vouch was questioning the idea that one should expect that sending a webmention means seeing it on the receiver's post.
#
jacky
I think doing that (setting hard expectations as to what's visible or not on one's site) would be a bit too encroaching
[jgmac1106] and [scojjac] joined the channel
#
[schmarty]
for sure! but then that also raises a big question for vouch which is, like: when do i as a sender need to go add a vouch to a webmention i want to send?
#
jacky
hmm okay
#
jacky
I think I want to capture my answer to that on the FAQ for the wiki
#
jacky
but I think _explicitly_ and _only_ when you notice that your Webmention hasn't been accepted _at all_ the first time around
#
jacky
the thing is: if you're trying to block someone - that could be a hint that you _are_ blocking them
#
[schmarty]
"i sent you a webmention and i don't see it on your site. i guess i'll try again by adding a vouch using <not-yet-existent UI>"
[calumryan] joined the channel
#
[schmarty]
ah like if the WM endpoint says "no thanks, i need a vouch for this" or similar?
#
jacky
hinting that a vouch is needed would be good
#
jacky
but I was thinking like if it kept spitting back a 400 despite one providing the source + target
#
[schmarty]
that's a missing piece of UI for a lot of folks who send webmentions, i think.
#
jacky
the failure case of a Webmention?
#
[schmarty]
failed outgoing webmentions as a thing to store and look at
#
jacky
ah yeah
#
jacky
(if anyone has ideas on when to do it - please add!)
#
jacky
[schmarty]: that kind of UX is something I never really considered b/c I would have the service do retries at a particular rate
#
jacky
but that's something to surface
#
[schmarty]
yeah, just one example of Vouch adding an implied human-in-the-loop where i would say most implementations are explicitly designed to be invisible.
#
[tantek]
pretty sure we had an HTTP error code for Vouch needed
#
[tantek]
that a site could respond to a webmention with
#
jacky
it's mentioning using Microsoft's 449
#
[tantek]
but yeah, since then I'm much more leaning towards rethinking the problem again from the perspective of questioning the social media assumption of "open responses" by default
#
[tantek]
yeah, that's the one jacky
KartikPrabhu joined the channel
#
[schmarty]
so that lets the sending system know a webmention won't be accepted without a vouch. but from there the /Vouch docs kind of drop into hand-waving territory
#
[schmarty]
> you can decide how much intermediate UI you want to show, "The comment you sent is awaiting some form of vouching that they link to that links to you" with a URL field to enter. Plenty of opportunity to crawl, cache, innovate there.
#
[tantek]
[schmarty] yes, that's deliberate because we literally had / have not figured out a "good" set of UIs and thus needed to (still!) encourage experimentation
#
jacky
cries in "we need a designer"
#
[tantek]
maybe we need to clarify that that's where we are with Vouch. it's definitely still something that "needs prototyping", "needs experimenting"
#
[tantek]
anyone coming to Vouch expecting "just tell me what to code" is going to get frustrated
marinin[t] joined the channel
#
[tantek]
because it's not are that level of maturity
#
[tantek]
at* that level
#
[schmarty]
tantek: for sure! i don't mean to sound like i hate Vouch, i just think it gets talked about a lot like it's "ready to implement".
#
[tantek]
jacky, maybe? mostly we need experimentation and iteration, even with clumsy / awkward designs
#
[tantek]
yeah we need to dispel that [schmarty]
#
jacky
more like for interfaces like the above mentioned
#
[schmarty]
one thing i love about these discussions of Vouch is how they shine a light on things that may be helpful to have in general, like encouraging webmention sending systems to have a way to surface stuff to you.
[chrisaldrich], KartikPrabhu and alex_ joined the channel
benwerd, slower_loris, alex_ and [KevinMarks] joined the channel
#
[KevinMarks]
Do we reframe vouch as a proposed way to construct an allow list?
#
[tantek]
KevinMarks, I'm not even sure burdening users with an "allow list" is the right approach/framing either!
marinin[t] joined the channel
#
[tantek]
I think we let Vouch remain for the historical attempt that was/is, and we redo those high-level use-case / social dynamics conversations
#
[tantek]
Understanding of "social media" systems and things like "respectful responses" has grown and innovated considerably since 2013, the context within which Vouch was proposed
jjuran, marinin[t], [tw2113_Slack_] and [Kai_Vong] joined the channel
#
[Kai_Vong]
has a publisher or well-known copywriter given the green light for using a http code as that ?
#
[Kai_Vong]
^http code^ensured parameter
[schmarty] joined the channel
#
jacky
[Kai_Vong]: which code? The 449?
#
jacky
can we change this assumption re: public-by-default with posts? https://indieweb.org/Micropub-extensions#Visibility
#
jacky
I know it says SHOULD
#
jacky
but I'm starting to move into the mindset that I'd tap to make this public by default (like if I syndicate the post then it can be considered public unless I also mark that the syndication target can support a protected or private visibility)
#
[Kai_Vong]
to be clearer; part of the question is around it being a legal keyword and does translate okay in other languages (such as French)dic ?
#
[Kai_Vong]
my keyboard is outdated ight now 😃 right now
jamietanna joined the channel
#
jamietanna
jacky I'd guess most folks are comfortable with the public-by-default, but I'd be interested to hear if there's a wider consensus against that
#
sebsel
do we know if most folks are comfortable with public-by-default?
#
aaronpk
jacky: it's written that way because it's the safe way to handle it
#
aaronpk
as in, it's unsafe if the client omits the parameter if the user wants the post to be unlisted
#
aaronpk
it's not about what the user wants, it's about what the app sends to the server. the app can default to unlisted posts in the UI, it would just always send that parameter in the micropub request
#
jacky
ah okay
#
jacky
and I guess we can pre-fill to understand from what's sent back in q=config (IIRC)
[barryf] joined the channel
#
jacky
okay then it's 'audience' I'm more concerned about
#
jacky
okay that's easy then (I can make some special internal URL that everything gets by default to make it private)
#
aaronpk
the concern is mainly interop in a safe by default way
#
aaronpk
adding a parameter means you need to absence of that parameter to work the same way as if the extension wasn't supported
#
jacky
right
#
[tantek]
hmm I thought for some reason that Bridgy Publish to GitHub handled skintone variants of the reacji it supported (e.g. 👍)
#
[tantek]
(it didn't and posted that as a comment instead of a reacji to a comment)
#
jacky
I guess that's me overloading 'visibility' as like "visible to who?" versus "how visible is this on my site when accessed via another means that is or isn't a direct link)
[snarfed] joined the channel
#
[snarfed]
PRs welcome though!
#
[tantek]
snarfed, indeed I should have remembered from the last time I touched that file 😂
#
[snarfed]
:troll:
#
[tantek]
snarfed, WDYT of expanding it to do so? as in, would you accept such a PR?
#
[snarfed]
sure! i haven’t followed github’s skin tone support, so i’d want some info and API doc links in code comments, but still definitely yes
#
[tantek]
I would do the skintone collapsing in granary
#
[tantek]
so you could thumb-up any tone you wanted on your site, and it would still "work" when POSSEd to FH
#
[snarfed]
oh interesting, ok, sure! for all four skin tone emoji? thumbs up/down, laugh, confused?
#
[snarfed]
that definitely sounds simpler
#
[tantek]
though it may only be the thumbs that are allowed
#
[snarfed]
allowed?
#
jacky
yeah IIRC confused doesn
#
jacky
confused doesn't have modifiers
#
jacky
unless I missed the memo lol
#
[tantek]
^ nor laugh
#
[tantek]
feel free to file an issue and cc me, got to run for now
[fluffy] joined the channel
#
[snarfed]
got it. thanks!
#
[snarfed]
another PR candidate might be removing most of the “just”s from bridgy’s docs 😆
Seirdy and vilhalmer joined the channel