#dev 2020-10-22

2020-10-22 UTC
[tw2113_Slack_], geoffo, nickodd, [KevinMarks], digiSal, [fluffy], treora, [tantek] and lahacker joined the channel; nickodd left the channel
#
Zegnat
[fluffy]: yeah, I have also been thinking about floating some sort of string value. But as [tantek] says, I do not like the ppronoun-nominative et al specifically because of the languages question.
#
Zegnat
I also wonder what I am marking up different cases/inflections for... computers? Do they really only need? Finnish has like 13 cases.
#
Zegnat
https://wiki.zegnat.net/microformats/pronoun has my thoughts and why I ended up with links
#
[fluffy]
yeah not wanting to get hyper-specific especially in an English-centric way is totally valid
#
[fluffy]
but I do feel that a p-attribute is much more important, given the exact same reasoning you have for a wiktionary link instead
#
Zegnat
All the way at the end of https://microformats.org/wiki/h-card-brainstorming#Pronouns there is also a note from sknebel wondering about publishing text-link-pairs as an option as a reaction to my links.
#
[fluffy]
like, if people see a familiar pronoun they’ll know how to use it. but the URL doesn’t necessarily convey the pronoun in an easily-readable way and it kinda gets in the way of human comprehension IMO.
#
[fluffy]
I think if you were to mark up your <a> with both u-pronoun and p-pronoun it’d be perfect
#
[fluffy]
My particular use case is wanting something to quickly show on a machine-passable profile page for people who have logged in to my site.
#
Zegnat
Hmm, that is also an interesting one, I would want to look at the output of u- and p- combined. But that might work without needing an extra wrapping h-
#
Zegnat
Although language support may be a reason to go with a wrapping h-pronoun
#
[fluffy]
language support is another entire mess of stuff though. I don’t think mf2py does anything with `lang` attributes.
swentel joined the channel
#
sknebel
if you don't want to look in detail into the h- you don't have to, just do it as a p- fallback value
#
[fluffy]
yeah, I just mean like <a class=“u-pronoun p-pronoun” href=“https://pronoun.is/she“>she</a> gives you both in a way that mf2 parsers Just Work. Or at least in a way that works with mf2py. I haven’t really worked with a lot of mf2 libraries.
#
Zegnat
<a class="p-pronoun h-pronoun" href="https://pronoun.is/she">she</a> also Just Works^{tm}. Will give you a pronoun property with both a name and url, and a fallback string value of she.
#
Zegnat
There are a bunch of options
#
[fluffy]
I don’t really understand how h-properties work, I guess
#
[tantek]
h-* is not a property, it's a root class name
#
[fluffy]
case in point
#
[tantek]
it creates a new object, into which you put properties
#
[fluffy]
right like… h-card/h-entry/etc. are objects that contain stuff, and I guess h-pronoun as an object that’s atatched to a thing in the same way that h-card is makes some sort of sense
#
[fluffy]
but to me it feels like a more natural fit as a property of the object, like, my pronouns are things about me, not things I own
#
[tantek]
it's not either or, it's both
#
[tantek]
that's the point of Zegnat's markup
#
[fluffy]
the stuff about h-pronoun on Zegnat’s markup seems to show why not to use it though
#
[tantek]
the company you work for is also a thing about you, and that company has its own properties, so it must also be its own object
#
[fluffy]
and the actually-in-use markup is just u-pronoun with no p-
#
[fluffy]
which is an okay way of providing links to pages about the pronouns but not to provide text to use as the link text
#
[fluffy]
(or present to the viewer)
#
[tantek]
why is it not ok to hyperlink the use of a pronoun to an explanation of it?
#
[tantek]
that would seem to be precisely the point of hyperlinks
#
[fluffy]
yeah I know, but what text do you show within that link? if the purpose is to show, like, a card catalog of people based on their parsed h-cards
#
[fluffy]
maybe it’s a limitation of mf2py but it doesn’t tell me any of the link text, it just gives me the URL of the link itself
#
Zegnat
This is what I was thinking with the wrapping h-*: http://php.microformats.io/?id=20201022070509472
#
[tantek]
"within that link"? what does that mean?
#
[tantek]
there is the linktext, and there is the destination of the link. which is "within"?
#
[fluffy]
I’m not sure if you’re confused about what I’m asking or if you’re being needlessly pedantic
#
Zegnat
The card of a person has one or more pronoun properties. Those pronoun properties have plain text values (he, they, in that example, in their value properties) but because of the use of h-pronoun you also have the ability to link together a plain text version *and* an explanatory link
#
[tantek]
a parser is not a crawler, it is not reasonable to say that a limitation of a parser is that it doesn't crawl URLs
#
Zegnat
Does that example make sense as far as h-* goes, [fluffy]?
#
[fluffy]
…I’m not saying that the parser needs to crawl the URL
#
[fluffy]
what I mean by link text is the text node inside the <a> itself
#
[fluffy]
like in <a href=“foo”>bar</a>, the bar
#
[fluffy]
the human-readable name/description of the thing being linked to
#
[tantek]
retry parsing Zegnat's example: <a class="p-pronoun h-pronoun" href="https://pronoun.is/she">she</a>
#
[fluffy]
yes and I don’t have any problem with things being marked up that way
#
[tantek]
the link text will be auto-implied as the p-name of the h-pronoun
#
[fluffy]
but that’s not how they’re being marked up
#
[tantek]
as well as the p-pronoun
#
[fluffy]
they’re being marked up as <a class=“u-pronoun” href=“foo”>bar</a> and there’s no p-pronoun or p-name or whatever
#
[tantek]
then it sounds like you've determined advantages/disadvantages of different methods of marking up pronouns
#
Zegnat
I think [tantek] is refering to the example I gave here in chat, rather than what I currently have on my site :)
#
[fluffy]
right, and I’m trying to understand why none of the commonly-used pronoun microformats are using a p-pronoun or the like
#
[fluffy]
yes, that example you linked to is a totally fine way of doing it. but it’s not how I see people actually doing it on their real live h-cards
#
[tantek]
probably because they're still experimental and haven't been consumed by any code that does anything with them
#
[fluffy]
I’m consuming pronouns on my site
#
Zegnat
Yeah, I do not know of any consumers
#
Zegnat
Oh, cool!
#
[fluffy]
or at least I try to 😛
#
[tantek]
there we go
#
[tantek]
that's the necessary feedback loop
#
[fluffy]
right now I consume p-pronouns and float it up to the UI on one’s logged-in profile (visible to the user and to the site admin)
#
[fluffy]
I also have made a change to Authl to fall back to p-pronoun if that’s available but p-pronouns is not
#
[tantek]
[fluffy]++ amazing
#
Loqi
[fluffy] has 20 karma in this channel over the last year (75 in all channels)
#
[tantek]
is that documented on the pronoun brainstorming page/section? that's a great thing to catpure
#
[tantek]
capture goodness sakes
#
[fluffy]
I should do that, yes
#
[fluffy]
I released this functionality in Authl back in early August
#
[tantek]
would it parse the <a class="p-pronoun h-pronoun" href="https://pronoun.is/she">she</a> markup correctly and be able to make sue of it?
#
[fluffy]
In theory, yes
#
[tantek]
argh, use* of it
#
[fluffy]
well, the current released versio of Authl wouldn’t, because it expects p-pronouns, not p-pronoun
#
[fluffy]
but I have, again, made a change to accept both property names
#
[fluffy]
If there’s more than one property it’ll just use the first one found. Because this stuff’s a mess.
#
[fluffy]
on Mastodon it just looks to see if any of the profile metadata slots has a ‘pronoun’ field and floats that up directly
#
[fluffy]
and I also don’t consume any HTML stuff because, you know, XSS/injection/etc.
#
[fluffy]
anyway you can see what the profile stuff gets about you with https://beesbuzz.biz/profile
#
[fluffy]
which reformats things in a friendly way and also includes some human-readable text about your access permissions 🙂
#
[fluffy]
I thought I had a screenshot of this somewhere on the publ site but I’m not finding it
#
[fluffy]
oh and it also provides a Bearer: token you can have your feed reader use to do authenticated feed access. I need to get around to wiring up TokenAuth too but there’s a huge chicken-egg problem right now.
#
[tantek]
yes would be great to capture at least a summary of which property variant(s) are consumed and what it does with them (how it uses the in display etc.) here: https://microformats.org/wiki/h-card-brainstorming#Pronouns
#
[fluffy]
I don’t have a microformats.org wiki account
#
[tantek]
hmm thought we solved that with the new wiki upgrade
#
[fluffy]
that’s why I haven’t documented my usage there
#
[fluffy]
I have a brief usage note on /pronouns on the indieweb wiki but I need to flesh it out
#
[tantek]
does it parse p-pronouns into separate words? like from "she/her/hers" into { "she", "her", "hers"} ?
#
[fluffy]
no, if there’s multiple p-pronouns then the first one wins
#
[fluffy]
I could change that though, this was just an expedient implementation
#
[fluffy]
my specific h-card format is: `<span class="p-pronoun p-pronouns"><span class="p-pronoun-nominative">she</span>/<span class="p-pronoun-oblique">her</span> or <span class="p-pronoun-nominative">they</span>/<span class="p-pronoun-oblique">them</span></span>`
#
[fluffy]
so the p-pronoun gets parsed as “she/her or they/them”
#
[tantek]
how much of that is required by the publisher? what's the minimal markup that would actually "work" in the consuming code to do something interesting?
#
[tantek]
required *of the publisher
#
[fluffy]
<span class=“p-pronoun”>she/her</span>
#
[tantek]
ok I'm confused I thought you said it only parses p-pronouns
#
[fluffy]
assume I made a typo
#
[fluffy]
or that I’m speaking from the future where I’ve released a version that parses p-pronoun or p-pronouns
#
[fluffy]
the main branch of Authl has that change, I just haven’t released it yet 😛
#
[tantek]
then I don't understand the use of both p-pronoun and p-pronouns and what functionality each adds
#
[fluffy]
I’m jsut trying to be compatible
#
[tantek]
it's premature to be "compatible" with experimental properties
#
[fluffy]
I feel like p-pronouns is the better name for it because there’s more than one, but other people are using p-pronoun
#
[tantek]
it's actually counterthetical to a good design
#
[fluffy]
well okay, I’ll settle on p-pronouns then 😛
#
[tantek]
experimental is the time to prefer good design and be much harsher about what is accepted and what is not
#
[tantek]
much *more opinionated* (rather than harsher)
#
[tantek]
because those implementation opinions help inform the design of the feature
#
[tantek]
to be clear, none of this is obvious, so please don't take this as any form of admonishment, more like humble clarification
#
[tantek]
incubation is still very much a developing art
#
[fluffy]
I also opted not to bother supporting pronouns on twitter because I didn’t want people to discover their pronouns were vi/vim or whatever
#
Zegnat
I always go for non-plural forms here, because I do not think any mf properties have ever been plural?
#
[tantek]
sure, good to start with pure indieweb publishing use-cases first
#
Zegnat
Will definitely test my site against beesbuzz.biz/profile, good to know of more h-card consumers!
#
[fluffy]
I think I only support pronouns on mastodon and indieweb
#
Zegnat
Does it parse Hogwarts Houses yet? :P
#
[tantek]
Zegnat, correct, mf2 property names are generally by design supposed to be singular, and then they can be multivalued
#
[tantek]
singular *names*
#
[fluffy]
[expletive] JK Rowling
#
Zegnat
[fluffy]: yeah, there is that, but I do not think I am removing my house from my site. Too many fond memories.
#
Zegnat
Of course future Martijn may decide differently
#
[tantek]
Zegnat, might as well just use tags (i.e. p-category) for that
#
[fluffy]
sure. but more seriously I’m keeping authl profiles simple and common basis sorts of things.
#
[tantek]
you don't need a whole property for a proprietary fictional scheme
#
Zegnat
[tantek]: but where would the fun in that be? ;) What if you wanted to parse specifically the magic wand length that was recommended to me? If you check my h-card parsed output it is all there
#
[fluffy]
and those are all just text strings, and it’s up to the consumer to be like “oh I guess homepage is a URL” or whatever
#
[fluffy]
because this is intended to be lightweight and human-readable, not an encyclopedia of semantic web whatever
#
Zegnat
But meh, all that is mostly extra cruft for fun when people test my h-card, not really meant to be anything but an easter egg
#
[tantek]
Zegnat, the fun is not not polluting property-space with a bunch of things that dilute the actual useful properties
#
[tantek]
s/not not/not
#
[fluffy]
I mean I dunno, I get what Zegnat’s going for here
#
Zegnat
They are all properly vendored as "pottermore"
#
[tantek]
for anything that's a simple text label, you don't need a property, that's the point. just use a category
#
[fluffy]
and I did notice the hogwarts house stuff when I was looking at your h-card for reference FWIW 🙂
#
[tantek]
Zegnat, it has a real cost unfortunately, that's the kind of reasoning that ends up with schema-org unnecessary complexity
#
[fluffy]
anyway the Authl profile fields are basically “twitter plus pronouns” since that felt like a pretty good set for its purposes
#
[tantek]
they've got whole object classes that should have been nothing but p-category tags
#
Zegnat
[fluffy]: I agree with the simplest output. That is often what you want for h-cards anyway, it is how I imagine digital rolodexes :) And I love getting the pronoun discussion rolling again.
#
Zegnat
I might revise mine to be h-* objects to see how I like it. That means parsers with experimental lang support (like the Typescript one) will actually get that information
#
[tantek]
the use / discussion of p-pronoun and p-pronouns is still confusing
#
[tantek]
it's not clear which a publisher should use and why (for what effect)
#
[tantek]
and the nominative / oblique stuff is very English specific so that feels like a poor design too
#
sknebel
Zegnat: hm, I somehow thought lang support wasn't experimental anymore, but I seem to misremember (and didn't catch the last few mf review sessions, so unclear on the state)
[chrisbergr] joined the channel
#
[chrisbergr]
Hogwards houses in h-card? I guess, the HP fans are more extreme than Trekkies!
#
[tantek]
sknebel, it was rejected in the most recent mf2 pop-up
gxt joined the channel
#
sknebel
k, need to go hunt for the notes from that then later
#
[tantek]
because of the way lang "inherits" etc. in HTML and not wanting to add that complexity to parsers
#
[tantek]
that's what I recall
#
Zegnat
Still behind a feature flag in some parsers, I quiet like it
#
[tantek]
I'm not entirely happy with that conclusion but that's where people seemed to have found consensus
#
sknebel
despite already aving implementations?
#
Zegnat
Been playing with the Typescript parser lately
#
Zegnat
And I find the implementation in that parser ticks almost all my needs, shrug. Although in reality, I am not consuming, so it is still very hard to say whether it fulfills its purpose or not
#
[tantek]
sknebel, well "implementation" is a bit strong. prototype of a proposal more like
#
[tantek]
and what was discovered/revealed during the pop-up was that the definition was insufficient to be useful to consuming code
#
[tantek]
therefore the same criticism of the existing prototypes
#
sknebel
k. I'll look at the notes/recording for details
#
sknebel
thanks
#
[tantek]
anyway like I said, I'd prefer we find some way to make "lang" work but I couldn't suggest a specific solution on the spot
#
Zegnat
Neither could I :(
#
Zegnat
But also, to reiterate: I do not have a consumer for it either, so I cannot make super strong arguments other than “I want something that differentiates my pronouns by language”
#
Zegnat
I would not be able to answer the question “differentiate them for whom?”
#
[fluffy]
oh on the topic of user profiles, has there been any work going forward on scope=profile for IndieAuth? Currently Authl requests that scope but doesn’t make use of any data that comes in from it, and instead just parses the h-card on the identity URL.
#
jamietanna[m]
Chrisbergr thanks! I can't take the credit though, I used a theme and just tweaked it for the stuff I wanted ☺
#
[fluffy]
I mostly just started requesting that scope to not have to deal with the spec change with scopeless requests
#
[chrisbergr]
The html lang attribute. Can this be used only on the html tag or could it also be used inline on certain microformat properties?
#
[tantek]
[fluffy] I added p-pronoun h-pronoun markup to my h-card on my home page (along with the human readable prose text obv), and I'm curious if that minimal markup is enough for your profile parser to see / do anything with
#
[fluffy]
[tantek] it would be except that the current released version of Authl only consumes p-pronouns 🙂
#
[tantek]
I don't understand what it means to consume "p-pronouns", like what values is it expecting?
#
[fluffy]
I mean as opposed to p-pronoun
#
[fluffy]
it literally just displays the value of the property
#
[tantek]
so the full literal string like "he/him"
#
[tantek]
a-ha ok
#
[tantek]
alright, worth experimenting with that too
#
[fluffy]
also authl only parses the h-card at time of login, and your last login was before my database got rebuilt, so you’ll have to log out and in again to get your profile to display
#
[fluffy]
my h-card is structured such that the containing p-pronouns has its stripped text appear like `she/her or they/them` with the individual pronouns themselves getting the p-pronoun-nominative or whatever
#
[fluffy]
(and I just ignore those properties)
#
[fluffy]
(in Authl I mean)
#
[fluffy]
with the current h-card on your site I would expect it to work
#
[fluffy]
ah I see you’e doing p-pronouns for the collected one and individual p-pronoun properties for the individual components. nice.
#
Zegnat
[fluffy]: re progress on scope=profile for IndieAuth, profile information is now fully integrated as part of the spec :tada:
#
[fluffy]
nice! is there a spec for what fields are expected and their meanings?
#
[tantek]
[fluffy]++ I logged out and logged back in and it looks like it worked!
#
[fluffy]
ah yes I see the profile on my admin page too 🙂
#
Zegnat
[fluffy]: yes, all part of the core IndieAuth spec now
#
[tantek]
this is some very nice kind microcopy: "When you logged in, your identity provider gave me some information about you. Here's the other stuff I know about you from that."
#
[tantek]
that profile page feels exceptionally welcoming
#
[fluffy]
looks like the image bridge is not working, or at least my message for it did not appear on the chat.indieweb.org log 🙂
#
[fluffy]
yeah I spent a lot of time on it 🙂
#
[tantek]
it shows
#
[fluffy]
I don’t remember if the Jinja template is part of what I released for my sample templates, let me check
#
[tantek]
well that's the first "improvement" I've made to my site in weeks, so thanks for the patience with the pronouns discussion
#
[fluffy]
of course that also means anyone who looks at that knows what the “I’ve seen you around” or whtaever text actually means in terms of permission groups 🙂
#
[tantek]
I was definitely encouraged with the knowledge that there was some code that did *something* with the property
#
[fluffy]
hm, I suppose I should be properly escaping the user properties in element attributes
#
[tantek]
feel free to document your consuming code current state (and aspirations) here: https://indieweb.org/pronoun#Implementation_Examples
#
[fluffy]
oh and the way publ templates work, you can actually get the profile as a card on any section of my site. so https://beesbuzz.biz/food/profile or whatever.
#
[tantek]
uh, I'm assuming you've voided that Bearer token 😛
#
[fluffy]
Whoops. I mean it’s not like you can see the whole thing anyway 🙂
#
[tantek]
that might be a good use for the 👁 icon to show it and have it hidden by default
#
[fluffy]
a <details> or whatever
#
[tantek]
yes with that markup 🙂
#
[chrisbergr]
I think about consuming with replies. Let's say I would like to have the info above my post: "I replied to a post by *tantek*. *He* wrote:"
#
[chrisbergr]
Should I always search for the 1st occurrence? Doesn't 'p-pronoun-nominative' make sense that I can really use the right one?
#
Zegnat
I would almost want to use <input type="password"> so browsers know not to cache the value or show it if your CSS or whatever does not load
#
[tantek]
depends on the language of the publisher vs the language of the prose you are publishing that makes use of the pronoun
#
[fluffy]
[Zegnat] but then it wouldn’t be copy-pastable into other things, or obvious what to do with it
#
[tantek]
"p-pronoun-nominative" is too English-specific to codify into a format which is why I'm choosing to not use it at all
#
[fluffy]
also there’s nothing to cache, it’s generated server-side at page load time
#
[tantek]
whether or not a pronoun is nominative etc. can be represented at the URL that defines the pronoun
#
Zegnat
Is copy always blocked on password fields? Maybe it is, was just an errant thought not something I tested
#
[tantek]
instead of burdening every publisher with understanding and getting nominative / oblique / etc. right in every language
#
[fluffy]
[chrisbergr] in that use case I’d just opt to format the quote text to not include pronouns at all, like “tantek wrote:”
#
[tantek]
sure, that's a reasonable approach
#
Zegnat
"too English-specific to codify" + 1
#
[tantek]
similarly, the URL for "he" could offer a <link rel="alternate" hreflang="de" href="/er" />
#
[chrisbergr]
Oh okay. I think german is too similar to this english specific thing that I didn't had that in my mind.
#
[fluffy]
anyway unfortunately I have no way to revoke individual tokens, I can only bump my session key and sign everyone out and invalidate ALL tokens, because currently I use Authl in signed-token mode (using itsdangerous, which is roughly equivalent to jwt)
#
[tantek]
so that the publisher also doesn't have to be burdened with figuring out the equivalents of pronouns in all human languages
#
[fluffy]
so, good thing that token isn’t actually valid as visible 🙂
#
[tantek]
Zegnat ^
#
[fluffy]
(so good thing I used /blog/profile, because the full thing is visible on other pages)
#
[chrisbergr]
My fallback would always be "they wrote".. So I guess I leave it at that 🙂
#
[fluffy]
anyway I already changed the markup to use <details>. also it’s not like bearer tokens can be used for anything useful right now.
#
[fluffy]
aside from viewing my private blog entries I mean. 🙂
#
[fluffy]
but maybe in 6 months I’ll have gotten around to implementing micropub or something, so hold on to that screenshot and try brute-forcing the rest of the token if you want to wreak havoc I guess
[fluffy] joined the channel
#
Zegnat
[chrisbergr]: German is slightly different from English I think? Has 4 cases (nominativ, genitiv, dativ, akkusativ). But you get into the really heavy stuff when you get to Finnish :P
#
swentel
how many are there in finish?
#
swentel
holy crap
#
Zegnat
Hmm, I counted 13 when I wrote my article, but now I count 12 there ...
#
Zegnat
Still, big yikes from me
#
[fluffy]
yeah also the concept of pronouns doesn’t really translate well to all languages
#
[fluffy]
like Chinese and Japanese both do completely different things from each other and from English
#
Zegnat
This is why I do not think separate mf properties for different inflections/cases scales. Even though as ComSci you might want to go that way specifically because it makes it easier to do “He wrote: …”
#
[fluffy]
IIRC Japanese basically only uses gendered pronouns for first-person speech, and most gender is implied by dialect and context.
#
Zegnat
[fluffy]: yeah, but I know too little about those languages to dare comment on them when I wrote my reponse to existing mf experiments
#
Zegnat
I think petermolnar said at some point that they do not really use pronouns at all in hungarian. (But may be misremembering.)
#
[fluffy]
and I forget how Chinese provides gender cues but I don’t think they even use “pronouns” per se
#
petermolnar
there are "personal" pronouns in Hungarian, but they are gender independent.
#
[fluffy]
but yeah really the only reason I did the p-pronoun-nominative stuff was because other people were doing it.
#
Zegnat
I wonder what we can do to skew less towards English language examples all over the wiki … I guess I am part of the problem as someone who mostly uses English online :(
#
[chrisbergr]
Zegnat: Of course in german there are more than in english. But you can translate the english ones 1:1, that's what I meant. The finnish stuff you linked shocked me 🙂
[capjamesg] joined the channel
#
[chrisbergr]
Anyways. There was a lot of informative stuff here in the last hours. Learned a lot. I'm glad that I was able to read all this.
petermolnar, jeremych_ and KartikPrabhu joined the channel
#
@DenverProphitJr
My automatic indieweb #webmention code is working nicely. Builds a JSON data file per interaction type. #jekyll parses json to display article comment cites. @DavidAmerland https://denverprophit.us/technical-seo/using-science-philosophy-organize-semantic-web.html https://twitter.com/DenverProphitJr/status/1318959342771703809
(twitter.com/_/status/1319234045394505728)
[KevinMarks], dckc, justus_, geoffo, lahacker, [Denver_Prophit] and [tantek] joined the channel
#
[Denver_Prophit]
[tantek] Webmentionable ring by topic stub. 😃
#
[tantek]
[Denver_Prophit] back in the day, Technorati literally had built what you're asking for, using rel=tag hyperlinks (before webmention)
#
[tantek]
that would be more a tagged directory of people 🙂
#
[tantek]
category links in h-entry could be used to build up a directory of posts
#
[Denver_Prophit]
But the h-card... the org section...
#
[tantek]
and for blogs, this might be a use-case for category links in h-feed
#
[tantek]
the org property is just a name, if you want to define details of an org, it needs its own h-card
#
[Denver_Prophit]
I did that in my h-card markup for the directory root.
#
[Denver_Prophit]
I went a step further, in jekyll, to create dynamic vcard downloads if someone else wants to contact me and include a bit more information about me.
#
[Denver_Prophit]
so back to the point... When I do article research and want to cite them... it would be great intel to know which sites can interact with a webmention
#
beko
without peeking into the source manually: a browser plugin could solve that. There are plents of such extentions in the wild. Even some nerdy ones. Like https://addons.mozilla.org/de/firefox/addon/gnu_terry_pratchett/ (never used it myself but I do notice the X-Clacks-Overhead *waves at jeremycherfas
#
[Denver_Prophit]
@beko That wouldn't help discover webmention-able by topic. So, back to the example, I want to write about airplane models. What blogs by topic support webmention.
[schmarty] joined the channel
#
[schmarty]
indieweb.xyz works a bit like this. It's organized by subtopics, and all content is posted to individual blogs, appearing on indieweb.xyz by virtue of being tagged on the individual post and by the originator sending webmentions
#
[Denver_Prophit]
[schmarty] I looked at that. Aggregated blog topics by a category tag. That's all I got.
#
[schmarty]
Each poster to indieweb.xyz is demonstrating support for at least outgoing webmentions
#
[Denver_Prophit]
[schmarty] https://indieweb.xyz/en/books is a broad taxonomy. IMHO
#
[Denver_Prophit]
right
#
[schmarty]
The taxonomy choices would appear to be up to the site's creator and probably doesn't match up with your expectations, but the mechanism seems like a good starting point if you do have a sense for your preferred topic ontology
#
beko
Frankly I see no usecase here. If I discover a blog during my own research that features Webmentions: Nice, I'm happy. Otherwise I don't care that my own Webmention to it will vanish in the server log. If I find myself interacting with a blog more often I ask it's author to add Webmentions. This works amazingly often.
#
[schmarty]
"Supports webmentions" is also a very broad and vague phrase
#
[Denver_Prophit]
@beko asking sites to "add" webmention support is difficult. Example: My published author friend uses Joomla. No support for webmention
#
beko
That's the price we pay for not having a product here but an idea :) Eventually some Joomla programmer picks this idea up and fixes this.
justus_ left the channel
#
beko
*bbl, food
#
[Denver_Prophit]
@beko right? I like webmention. This is more of a brand advocate call to make it more popular. Videos I see are 2008~2018 like [tantek] If there's no demand and popularity there's no reason for someone to add it. I thought it would be nice to discover a catalog of webmentionable sites by keyword search. (i.e. airplane modeling). That's my two-cents 😃
#
[Denver_Prophit]
It's beneficial when I blog per se about model airplanes that other bloggers can find my article and interact/reply/favorite and get a link back. The fate of pingback still stands with aweful spamming.
#
[Denver_Prophit]
be back soon!
#
[tantek]
what was I going to say is we could use webmention itself to build up a directory like that
#
[tantek]
Hey all y'all IndieWeb developers (yes including WordPress GWG 🙂 ) — next week is W3C's TPAC breakout week (kind of like a week long virtual BarCamp), and for the first time the general public is invited to participate. You can register and see the schedule of sessions here: https://www.w3.org/2020/10/TPAC/public-breakouts.html
#
GWG
[tantek]: What sort of sessions usually come up?
#
[tantek]
GWG, here is the full list of sessions! (all proposed by W3C participants in the previous two weeks) https://www.w3.org/2020/10/TPAC/breakout-schedule.html#calendar
#
GWG
I'll see if something tickles my fancy
#
aaronpk
i'll try to join that one. sam did a session on that at IIW yesterday which went pretty well
#
[tantek]
GWG, you may be interested in https://www.w3.org/2020/10/TPAC/breakout-schedule.html#ResGeo (and aaronpk too)
#
[tantek]
I'm facilitating this session, and definitely encourage folks here who are particularly values/principles-opinionated (per /principles) to participate! https://www.w3.org/2020/10/TPAC/breakout-schedule.html#focusvalues
[KevinMarks] joined the channel
#
[KevinMarks]
If you try to send webmentions through mention.tech it keeps a record of them as well. I need to beef that up a bit so it is a useful way to backfill them.
[fluffy] joined the channel
#
mayakate[m]
Michael Lewis is doing cool things with https://searchmysite.net/search/browse/ I wonder if it might be possible for it to feature as a selector whether a site advertises a webmention endpoint.
#
[schmarty]
> we are updating the platform to move from a Wiki approach with the content in a MySQL database, to a JAMStack approach with the content being hosted in a Git repository
KartikPrabhu joined the channel
#
[tantek]
that's good right? I feel git repo is a better backend than MySQL
#
Zegnat
Not having read the blog post yet, my first fear would be searchability of file storage vs database. But maybe not an issue.
#
[tantek]
lol MySQL index for search is crap anyway. any "real" search built into apps built on MySQL is done with additional actual indexing
#
[tantek]
point being, either way you have to do explicit work to build a usable / useful search index on top of your storage layer
[chrisbergr] and [KevinMarks] joined the channel
#
[KevinMarks]
Up to a point - mdn shouldn't be that much text
#
sknebel
that does also seem to mean wiki editing only through github ( with PRs?)? That's a pretty big shift - not necessarily a bad one depending on who actually is editing
Lionir joined the channel
#
Lionir
Does Monocle not support p-summary? I would have e-content but it's not shown on the homepage and don't feel like building a seperate page for it. I wish I could just say the e-content is where the h-entry is linked but that doesn't seem to be a thing (and it'd probably not be ideal for readers anyhow).
#
aaronpk
i think it does only for certain things? I can't remember now actually
#
Lionir
It's in a h-entry, you can see here : https://thelion.website/
#
aaronpk
i don't see it on the preview page so i guess not! https://monocle.p3k.io/preview?url=https%3A%2F%2Fthelion.website%2F
#
aaronpk
wouold be a good feature to add tho
#
Lionir
Yeah, I tested there hence why I asked, haha.
#
Lionir
I was wondering if I had just messed up the formatting somehow.
#
aaronpk
looks good
#
aaronpk
only thing is the dates should be ISO8601 instead of english
#
aaronpk
monocle is happy parsing them but not everything might be so generous :)
#
jacky
that lil' progress bar before it renders is nice!
#
Lionir
Does `date(format="%Y-%m-%d %H:%M", timezone="America/New_York")` look like ISO8601? Honestly, these things go over my head, haha.
#
Lionir
No, it's not. Hmm
#
Lionir
Actually, no, that seems right according to https://tera.netlify.app/docs/#date
#
sebbu
iso needs seconds and timezone too
#
Lionir
I think this works then `date(format="%Y-%m-%dT%H:%M:00Z")`
#
Zegnat
Also, dates should probably be the WHATWG subset rather than actual ISO, if we are talking HTML time element / datetime attribute. (I know, it is confusing.)
#
Zegnat
Lionir: there are a bunch of examples here if you just want to compare values: https://html.spec.whatwg.org/multipage/text-level-semantics.html#datetime-value
#
Lionir
Time is so complicated. I think I'll use datetime=date(format="%Y-%m-%dT%H:%M:00Z") and the shown part be date(format="%Y-%m-%d")
#
Lionir
Nope, that'd be wrong, datetime seems to not want hours or anything like that.
#
Lionir
I want a epoch attribute to stop this pain, aha
#
Lionir
According to wikipedia, `2020-10-22` this is valid ISO8601 and WHATWG seems to accept that too.
#
Lionir
I'll just do that, I don't plan to post more than once a day so it's unimportant, I don't even write the time of day they are published so yep, gonna do that.
#
aaronpk
yeah `2020-10-22` is fine
#
Lionir
Perfect.
#
Lionir
Btw, should I make an issue on Github for monocle p-summary support?
[tantek] joined the channel
#
[tantek]
spent way too much time on those WHATWG date & time details
#
[tantek]
tl;dr: ISO8601 has A LOT of valid syntaxes, the W3C Date-time note (and IETF RFC on dates) has a really *dumb* subset (like please do not use that), and it took a lot of work to come up with an actual user-informed and usability-informed subset that we dropped into HTML5
#
[tantek]
also Atom uses the IETF RFC subset on datestimes so those errors in design propagated into Atom
[chrisbergr] and maxwelljoslyn joined the channel
#
maxwelljoslyn
scrolling back through the chat, I saw questions about (third-person) pronouns in Chinese. short answer: he/she/it are pronounced the same ("ta") but written with 3 different characters; rising awareness of gender neutrality has led some people to adopt a "they" equivalent spelled by simply writing out "TA"
#
maxwelljoslyn
and there is no inflection or marking for case in Chinese, so he/she/they are just ta/ta/ta - there's no equivalent of e.g. "him" "them"
#
maxwelljoslyn
and there is no inflection or marking for case in Chinese, so he/she/they are just ta/ta/ta - there's no equivalent of e.g. "him" "them"
#
maxwelljoslyn
that is to say, "he" and "him" are both "ta", no change.
krychu and joshproehl joined the channel