#dev 2018-12-17

2018-12-17 UTC
snarfed, iasai, [kim_landwehr] and [eddie] joined the channel
#
[eddie]
!tell manton There is a discussion in the IndieWeb channel about an error with a Micro.blog site and Aperture. https://chat.indieweb.org/2018-12-16#t1544997797573400 The error seems to say that the Micro.blog token is returning http rather than https, I'm wondering if that needs to be updated on M.b IndieAuth
#
Loqi
Ok, I'll tell them that when I see them next
#
Loqi
[[kim_landwehr]] Monocle
iasai, clintpatty and [eddie] joined the channel
#
GWG
[eddie]: The NWS is now an optional weather provider for my plugin, but it isn't ready to deploy yet, so I can't post on it
iasai, cweiske, snarfed, swentel, [Marco], TripFandango, trip_, jgmac1106 and [jgmac1106] joined the channel
#
swentel
hmm how should one handle the photos in the feed from grantcodes[m] ?
#
swentel
mf2 parsing lists urls pointing to the original/big image, so should there be an additional request then to get that image in a reader for instance ?
#
sknebel
not sure I understand the question
#
swentel
yeah, bit cryptic I guess
#
swentel
let me try like this
#
swentel
in the photo array, there are a bunch of photos
#
swentel
but they point to a url, not a source image
#
swentel
so, in my reader, I don't see images because I assume they are images, not web pages
#
sknebel
yes, that seems wrong on the sites part
#
swentel
so I wonder how aperture handles this for instance
#
swentel
because xray gives the same result (and I think in that setup xray fetches the feeds)
#
sknebel
it probably doesn't
[Vincent] joined the channel
#
sknebel
photo property is supposed to point to image files
#
swentel
that's what I assumed too
#
swentel
'Be sure to markup your actual photo image (or hyperlink to higher quality image) class="u-photo"'
#
swentel
that's on https://indieweb.org/photo though, so maybe confusing ?
#
sknebel
(the core problem probably is that nobody has done a concerted effort to get galleries figured out and written down markup-wise)
#
swentel
oh, I see yes
#
sknebel
although e.g. a multi-photo post (with photo pointing to images) somewhat "works"
#
swentel
yeah, most h-feeds I follow are fine
#
swentel
I understand why he links though, the thumbnail photos are all scaled to 200x200
#
sknebel
sure, could link (hidden then) to the fullsize images though
iasai, jgmac1106, barpthewire, eli_oat, swentel and [kevinmarks] joined the channel
#
[jgmac1106]
Also seems like a reasonable expectation for a photo collection or gallery to have thumbnails
#
[jgmac1106]
Collections as well... I was looking for guidance on or in the wild examples
#
[jgmac1106]
Starting to see them on pixelfed and micro.blog so might be something to put some thought into
#
Loqi
[Greg McVerry] But this gave me a cool idea... Could I wrap each section in a flexbox and make each <p> tag a grid with flex-grow set to one.. But then scrape my citation counts from Google Scholar and dynamically have flex-grow change in proportion of % of total c...
[grantcodes] joined the channel
#
kisik21
good morning IndieWeb. Is it possible to make an accordion-like spoiler without JS
#
[grantcodes]
kisik21: details + summary elements
#
kisik21
[grantcodes]: sorry I don't understand what do you mean
#
kisik21
you want me to tell about my use case?
[tantek] joined the channel
#
kisik21
I'm experimenting with markup and I want to make CW spoilers on my site like mastodon, but being sort of a JS hater I'd prefer to have it work without JS enabled since it is an essential feature of my site
#
[tantek]
Yes. See HTML summary and details elements
#
[tantek]
Ah what grant codes said :)
#
[tantek]
We need a good code example of those on /content_warning
#
kisik21
woah, wait, HTML HAS THIS BUILT IN?!
#
kisik21
Sorry I'm a bit shocked
snarfed joined the channel
#
[jgmac1106]
HAD exact same question... I use bootstrap accordion tables at work... Wantes to know if I could use collapse with no JS....
#
[jgmac1106]
So @kisik21 my professional currency is how many citations I get... Had this random idea after trying to learn flex-grow... For prettier tables....
#
[jgmac1106]
Could I have a default flex-grow set to 1, but then that number changes based on citation count... Be fun to trigger fonts.... This would take JS
#
@m4d_z
@nhoizey tiens, à l'occaz, je veux bien causer #webmention avec toi ;) (lors de notre soirée geekeries par exemple)
(twitter.com/_/status/1074672550263877633)
#
[jgmac1106]
... Hey wait... Could you use flex-grow on click, for content warning? Maybe change a z-index and hide a box? Have a box set to column.. Stack it? Use opacity?
[Vincent] joined the channel
#
[Vincent]
I would like to see something “like” a spoiler MF tag. Being able to caveat spoilers or sensitive content, would mean things like search engines etc wouldn’t bring back potentially harmful (in whatever manner) data. It could also aid in things like data sensitivity when extracting excerpts from one site to another, such as comments.
#
[jgmac1106]
But @kisik21 actually pulled over on side of the road to learn about HTML accordion...
#
kisik21
but the summary and details is exactly what I need
#
kisik21
so case solved
barpthewire, [eddie], iasai, swentel, snarfed and Tin2 joined the channel
#
[jgmac1106]
[sknebel] [tantek] [gRegorLove] do you see a collection as an h-feed embedded in an h-entry?
#
[jgmac1106]
There was u-repost-of h-cite, could see that and a p-category with rel=prev and rel=next for navigation
#
[jgmac1106]
Meant to cc [aaronpk] @zegnat [grantcodes] and anyone else who might care
#
[eddie]
jgmac1106 I believe when it's been previously discussed, a discussion is an h-entry with child h-entries
#
[eddie]
no h-feed needed
#
[jgmac1106]
Perfect... Not noted on wiki... Could use some in the wild examples... Thinking of doing that with my vita....
#
[jgmac1106]
Do I add a p-category on top level h-entry or does each h-entry need a p-category tag?
#
[jgmac1106]
Have you looked how [manton] and [cleverdevil] are marking up galleries
iasai joined the channel
#
@leekelleher
↩️ Same for me. I've been looking at some kind of Webmention + Bridgy set-up. https://indieweb.org/Webmention#Bridgy (I haven't got far with an integration yet)
(twitter.com/_/status/1074693182674399232)
#
[eddie]
What is collections?
#
Loqi
A collection is a type of post that explicitly lists and/or embeds multiple other posts chosen by the author https://indieweb.org/collections
[grantcodes] joined the channel
#
[grantcodes]
Collections are kinda complicated. I store mine as children on the collection post, but then since I mainly just use photo galleries and to get it to work in readers I show preview photos on the collection marked up with u-photo, although it might not be technically correct it works really well
#
[eddie]
I will say, your photos don't appear in my reader 🙂
#
sknebel
right, that triggered the entire conversation :D
[cleverdevil] joined the channel
#
swentel
(same) hehe
#
[cleverdevil]
I don’t like how I’m currently doing it, but it works pretty reliably because it’s so simple. Basically, Known supports multiple photos on a single “photo” post. Plus side: works easily. Minus side: you can’t interact with individual photos.
#
[eddie]
[cleverdevil] So you're essentially doing multiphoto posts as opposed to collection posts
#
[cleverdevil]
Precisely.
#
[eddie]
:thumbsup:
#
[grantcodes]
Yeah not sure what's wrong with my photos in together, they get picked up by aperture
#
sknebel
really?
#
swentel
really ? then aperture follows the link ?
#
sknebel
you have u-photo properties that point to non-image files, I'd not expect that to work
#
[grantcodes]
Oh nah there's something funny going on, might have broke my mf2 at some point
#
[eddie]
ahhh maybe that's it 🙂
#
[eddie]
Tricky mf2, that's happened to me before as well
#
sknebel
what's a bit annoying (IMHO) with a nested object for the photos is that if you'd want it to work with the "value" of the nested object (which IMHO it should), you need to put the url of the image file as u-url of the nested object
#
sknebel
which means you can't just nest a h-cite of a photo post
#
[grantcodes]
Yeah I changed links somewhere so it is picking up the links to the photo posts not the urls of the images
#
sknebel
not sure how to resolve that except by having an explicit rule to *not* use the value, which is also weird
#
[grantcodes]
sknebel I think that is my exact case that has broken them. Not sure I fully understand though
#
sknebel
I don't think its exactly the same, sicne I didn't see nested objects in the mf2 output
#
sknebel
unless you store nested objects internally
#
[grantcodes]
Yes and no 😛
#
[grantcodes]
My storage is pure mf2 json. And the collection has an array of children with links to the child posts, and then it is reassembled with nested objects when requested
#
[grantcodes]
But I don't output them as nested in the html, because I only want to show a preview image, not the date & title & location and everything else that could be there
#
[eddie]
jgmac1106 thanks for pointing out the info wasn't on the wiki. I tried to add some basic info based on stuff that has previously been discussed in chat
#
sknebel
[eddie]: the examples at the top of the page by aaron and grant do not follow that though?
#
[grantcodes]
Well my markup currently doesn't really work, but follows that idea
#
[grantcodes]
<div class="h-entry">
#
[grantcodes]
<h1 class="p-name">Collection</h1>
#
[grantcodes]
<div class="gallery">
#
[grantcodes]
<img src="https://thumbnail.com/small-image.jpg" />
#
[grantcodes]
<a class="u-photo" href="https://urltophotopost.com">
#
[eddie]
I thought aaronpk's did last time I looked
#
[grantcodes]
That's a simplification of my current markup but now I'm not sure how to get to to work how I want with mf2
#
[eddie]
hmmm let me look again
#
[jgmac1106]
I think okay to have in the Wild examples and brainstormming below
#
[eddie]
That's strange, because aaronpk is the one that I discussed this with
#
[eddie]
his thinking must have changed after he made that post?
#
[eddie]
since it was in 2014
#
[eddie]
strange
#
[jgmac1106]
Is the thumbnail an u-repost h-cite of full img [grantcodes]? or is it better to have a <link> to full resolution?
#
[jgmac1106]
u-syndication... Not sure anyone uses u-repost
#
[jgmac1106]
Or is a thumbnail a kind of p-summary? A condensed version of a full source?
#
[grantcodes]
The link is to a full post, the image is a thumbnail, but should send only the full res image src to the reader
#
[eddie]
😞 indiechat.search.cweiske.de is still not working. So I can't search to find my previous discussion in chat about collections
#
[jgmac1106]
Oooh no.... Love the search
#
sknebel
[grantcodes]: easiest fix would be adding a hidden link to the full-res image then I guess. Might also be a potential use case for the srcset support that was suggested
#
sknebel
(but that isn't in yet, since nobody was willing to make any comment on it last I asked)
#
[grantcodes]
Yes unfortunately can't see any option apart from hidden markup somewhere
#
[jgmac1106]
.... Thumbnail as background img in css
#
[jgmac1106]
Make a div as a link. The div has a background img set your thumbnail but links to full u-photo
#
sknebel
but the clickable link should go to the photo post page, not the image file
#
sknebel
and please do not use div background imgs for content
#
[jgmac1106]
Didn't think a thumbnail was content as a copy of content. But yeah keep design and content different
#
[jgmac1106]
. Aaah that's why you want hidden link... So readers get full img
#
[cleverdevil]
I currently do the div background-image thing (well, Known does), and I definitely wish I could accomplish the same using img elements instead.
#
[cleverdevil]
I am sure its possible, but I haven't had a chance to get around to it.
#
[jgmac1106]
Didn't realize Known did...knew it squashed pics and linked to full version
#
sknebel
if what is possible?
#
[jgmac1106]
To not use img tags and not background imgs in a div for img placeholders... Didn't know Known did that
#
[eddie]
I think essentially you would do an img with low z-index in css
#
sknebel
oh, misread it as "not sure it is possible", nvm
#
[jgmac1106]
And I meant "to use" just img tags so what counts as hidden data a <link> tag or z-index:-10
#
swentel
hmm but both images in source will still let the browser load them, not very friendly either no ?
#
[cleverdevil]
I am auto-cropping to square.
#
[jgmac1106]
Almost like you need a u-uid or something like it for a full resolution and the u-photo for the cropped of thumbnail
#
sknebel
the thumbnail image isn't cropped, the page isn't showing it fully
#
KartikPrabhu
why not always link to the full version and the reader/consumer can do whatever lower resolution processing it wants?
#
KartikPrabhu
that is what I do internally on my dite
#
[grantcodes]
Is there a way to prevent microformats picking up an img as implied photo?
#
sknebel
have any u- property on the object
#
[grantcodes]
But if it is not a property? I have to add a junk one?
#
sknebel
(well, parsers might not do that yet)
#
sknebel
on the object, not on the img
#
KartikPrabhu
that seems like a bad hack. If there is a good use-case maybe file an issue with mf2 parsing spec
#
sknebel
[cleverdevil]: I think in modern browsers you can get that cropping effect with CSS object-fit on img tags
#
sknebel
[grantcodes]: what's the specific example?
#
[grantcodes]
Use case is thumbnails, I have very small ones and only want the full res ones in the parsed mf2
snarfed and [tantek] joined the channel
#
[tantek]
Indeed, could you provide the h-* markup snippet and context?
snarfed joined the channel
#
sknebel
huh? but then you do have a photo property on the object?
#
sknebel
then there is nothing implied
snarfed, iasai and jgmac1106 joined the channel
#
[grantcodes]
It almost maps to my internal mf2 apart from it picks up the thumbnail as the photo on the child
#
sknebel
yeah, I'm not convinced of the child, especially combined with u-photo on the top level
snarfed joined the channel
#
[grantcodes]
The u-photo is a hack really, since I don't think there are really many consumers of children
#
[jgmac1106]
Chronus did it.
#
sknebel
hm, the implied photo reduction change hasn't actually been added to the spec yet, and not even merged in the parsers (I somehow thought that was done), so that needs doing too
#
sknebel
so yeah, tricky to avoid here without adding something to break it intentionally or repeating something
jgmac1106_ joined the channel
#
swentel
so maybe a <data> element is the cleanest solution for the original problem ?
#
sknebel
(I'm also not convinced the children thing works well - e.g. is it specific to h-entries? shouldn't it be h-cites? what about h-cites for non-collection purposes, or h-cards?)
#
sknebel
swentel: yeah, to just add the full urls to a photo-property without showing the img, put it in a data tag
#
[jgmac1106]
[sknebel] I would argue a a child h-entry is only an h-cite if a post with a url exists...but could I have an h-entry in a collection that doesn't live anywhere else
#
[grantcodes]
I don't think they are citations though, I'm not referencing the photo, it is an integral part of the collection. I think a h-entry with a nested h-cite should make sense without the cite. A collection with no child(ren) is not of use
#
sknebel
you are referencing hte photo post - if you aren't, why is it linked?
#
sknebel
you are also only showing some subset of it's properties
#
sknebel
both could be arguments for h-cite IMHO
#
[grantcodes]
ok but I think either way I still have the same issues though? Or if it is a h-cite does that tell parsers they should check the url for the full content?
#
[tantek]
sknebel re: implied photo reduction change hasn't actually been added to the spec yet, can you at least cite the GitHub issue? It's hard to follow-up when not knowing exactly what needs follow-up
#
Loqi
[aaronpk] #21 reduce instances where photo is implied
#
[grantcodes]
Hmm this is getting close to what I want, sort of
#
sknebel
[grantcodes]: parsers don't work on that level, was more "are there obvious problems with the proposed approach"
j12t and jgmac1106_ joined the channel
#
[jgmac1106]
FYI [cleverdevil] no nesting of any h* in Known...every time I have tried it gets stripped away and leaves just the h-entry
[cleverdevil] joined the channel
#
[cleverdevil]
Strange... that doesn't seem like something Known should be doing.
#
[cleverdevil]
In fact, I don't think it does any sort of processing of the HTML that is produced within a template, for example.
#
[cleverdevil]
So, if you're working on a plugin that wants to do that, it should work based upon my experience/understanding.
#
[cleverdevil]
If you're trying to put raw HTML into the editor in the web interface, all sorts of wonky things happen because rich text editors are hard 🙂
#
[cleverdevil]
(In that case, publish your HTML via Micropub instead, which will bypass the editor entirely).
iasai joined the channel
#
[tantek]
[jgmac1106] FWIW you really should make sure you have a good consuming use-case before publishing *any* nested h-* structures, because they are non-trivial, and thus easy to get wrong, and thus if you're not *testing* them with a consuming application you may be doing more harm than good by publishing
#
[cleverdevil]
(That too ^^^)
#
[tantek]
e.g. we have lots of consuming code for h-entry > p-author h-card, so that's well understood and testable
#
[tantek]
throwing a bunch of random markup out there that you *think* is right but have no way of testing the effects of is more harmful in the long run than helpful
#
[tantek]
you end up creating more backcompat mess for future consuming applications, means more work for them, means less time they're spending on actual user features
#
[tantek]
(same thing with new CSS features etc.)
snarfed joined the channel
#
jgmac1106
[just been reading the h-resume….has so many nested h* items
#
jgmac1106
my personal use case for consuming is really around rubrics and just nested h-reviews where the ratings would total in h-review-aggregate…but building the grading system long off….need to get all the important pieces in place first
#
[tantek]
what is a rubric
#
Loqi
It looks like we don't have a page for "rubric" yet. Would you like to create it? (Or just say "rubric is ____", a sentence describing the term)
jackjamieson joined the channel
#
[tantek]
yes an h-resume consumer would be a good thing to build to test / validate h-resumes
#
jgmac1106
I also thought I would like to follow people’s vitas/resume if they updated every publication as a h-entry…but probably better for now to publish a notification of each publication and then an h-cite on a vita page
#
[tantek]
at least in that instance we have some experience from past hResume consumers that as long as we stick to a similar enough structure we can have some confidence that it will work
#
[tantek]
what is a vita?
#
Loqi
It looks like we don't have a page for "vita" yet. Would you like to create it?_ (Or just say "vita is ____", a sentence describing the term)
#
[tantek]
[jgmac1106] having trouble following your examples a bit - sounds like you may be using some academic-specific terms?
#
jgmac1106
Can I redirect vita to resume from here or do you go to wiki
#
[tantek]
is there a difference?
#
jgmac1106
one is longer and more useless than the other
snarfed joined the channel
#
[tantek]
or is vita like "c.v."? similar but not same as resume, more european than american?
#
jgmac1106
I didn’t know how to mark up publications and presentations using accepted h-resume so I left them blank: https://jgmac1106homepage.glitch.me/resume.html
#
jgmac1106
(haven’t done my media queries yet)…..but I want to follow @kartikPrabhu’s vita to knwo when ever he publishes a black hole paper….To do that each publication would end up a nested h-entry…
#
jgmac1106
more academic c.v. ois curriculum vita or curriculum vitae….depending on which side of the bed your rolled out of
#
jgmac1106
rubric is a measurement tool to assign ratings to specific critierion. They can be holisitic where multiple criterion are measured on a single scale, or criterion based where the rubric consists of aggregated ratings with each criterion judged on the same scale, or domain based where aggregated scales across multiple indicators consist of criterion and a degreee of quality scale. Rubrics, consisting or ordinal data, are often, many argue incorrect
#
jgmac1106
used as numerical data in theoretical models or personnel decisions such as human resources.
snarfed joined the channel
#
jgmac1106
vita is (vitae, C.V.) a specialized [[resume]] for academics that also include publications, presentations, grants, service, and creative activity not usually included in a traditional resume.
#
jgmac1106
Known doesn’t have the input fields for manual webmentions on the webmention endpoint and I am getting this error trying telegraph: https://telegraph.p3k.io/webmention/12gUsvvhkTVUMGo4kC/details am I doing something wrong?
[adamprocter] joined the channel
#
[adamprocter]
comments and thoughts very welcome]
#
jgmac1106
awesome [adamprocter] though we might want to move to the #IndieWeb channel for the comments unless dev specific questions about the draft
#
jgmac1106
you can use chat…just know it is off record and not archiveds
#
jgmac1106
…but we will take all the help i can get to build the headless IndieWeb CMS…..so many pieces falling into plaxce
#
jgmac1106
you might want to read through the methods here…describe how I want to build and use parsers and machine learning for bots: https://docs.google.com/document/d/1OrwDAzNPIn1DbJyUxbvyqVciYZUAgjO7dstQ7FmiGQQ/edit?usp=sharing
#
GWG
Afternoon
iasai joined the channel
gRegorLove, [jgmac1106], eli_oat and [pfefferle] joined the channel
#
sknebel
!tell [snarfed] made a PyPI account (sknebel), can you add that to mf2py or do I have to ask someone else?
#
Loqi
Ok, I'll tell them that when I see them next
leg, [chrisaldrich], barnabywalters, iasai, cjwillcock, [tantek] and [Vincent] joined the channel
#
jacky
does anyone have a webmention processing implementation that's JSON-friendly
#
aaronpk
what does that mean?
#
jacky
like could one pragmatically inquire about the status of a webmention that they've sent
[eddie] joined the channel
#
aaronpk
webmention.io has lots of that kind of thing
#
jacky
like if I sent one to you from my site but I'm not using webmention.io
#
jacky
lol yeah that's what I was trying to avoid
#
jacky
eh okay
#
aaronpk
I don't understand the question then
#
jacky
I guess it's a bit of sugar then
#
jacky
nah like let's say you handled webmentions as part of p3k but not via webmention and I had it from jacky.wtf as a in-built feature
#
jacky
I want to take the value of the location header and read its body to determine if my webmention was processed successfully
#
jacky
but with no user intervention
#
jacky
clicks
#
jacky
ooh interesting
#
aaronpk
telegraph does some of that too
#
jacky
I think that works for me
#
jacky
thanks!
#
jacky
ooh there's a callback option there
#
jacky
I think I like that one the most
#
aaronpk
that page is mostly brainstorming, but feel free to implement something and document it
#
aaronpk
I think that's the next thing we need to standardize as a webmention extension
#
jacky
definitely
[jgmac1106] joined the channel
#
[tantek]
!tell [jgmac1106] my search for openwashing articles found me this one https://blogs.plos.org/absolutely-maybe/2017/08/29/bias-in-open-science-advocacy-the-case-of-article-badges-for-data-sharing/ which appears to have some good criticisms of open badges efforts
#
Loqi
Ok, I'll tell them that when I see them next
#
[jgmac1106]
Thx [tantek] yes the #OpenScience team at MoFo developed an entire badging system for academic publishing... Not sure it was ever used
[cleverdevil] joined the channel
#
sknebel
aaronpk: in case you missed it last night, kim_landwehrs error screenshot came through and we were stumped - ideas? https://chat.indieweb.org/2018-12-16#t1545004083047300
jgmac1106 joined the channel
#
[tantek]
[jgmac1106] and did they include mocking or critical badges?
[matpacker] joined the channel
#
[tantek]
like the now summary image at the top right of: https://indieweb.org/openwashing 🙂
#
[jgmac1106]
No I didn't read article you posted.... Reason in the ooenbadges directory I am labeled as "Community Skeptic"
iasai joined the channel
#
[tantek]
so did they award you a "Community Skeptic" badge?
#
[tantek]
looks like about half that didn't make it into the page - may have edit /rubric manually and add it
leg, [kevinmarks] and [chrisaldrich] joined the channel
#
[jgmac1106]
No... Just a page in a slide deck... Open badges community issue badges... Funny... Stats had us over a million... But I never met anyone that had them
#
[jgmac1106]
Beyond all the blog ranking badges (which have nothing to do with spec) love to display
jgmac1106 joined the channel
#
[tantek]
yes I'm pretty sure the stats were all inflated
iasai joined the channel
#
aaronpk
sknebel: that error is saying that the token endpoint reports the URL as being "http" but when they logged in the URL was "https"
#
sknebel
aaronpk: so issue with micro.blog?
#
aaronpk
yeah looks like it
jgmac1106, iasai and snarfed joined the channel