#dev 2020-02-08

2020-02-08 UTC
#
[KevinMarks]
As was predicted in the OAuth 2 spec
#
aaronpk
haha yes
#
aaronpk
who could have seen this coming
#
[KevinMarks]
[fluffy] one way to do the cite is with a fragmention link to what you are quoting, though that works better when it's part of a longer post.
#
[KevinMarks]
But that gives you the "preserve what I was quoting" part of the problem
[fluffy] joined the channel
#
[fluffy]
[KevinMarks] yeah that’s not the use case I’m talking about though
#
[fluffy]
I’m not “quoting” a thing, I’m propagating a link that someone else shared.
#
[fluffy]
But I don’t want to share their share to the link.
#
[fluffy]
Because that is both bad UX and also leads to link rot.
#
[tantek]
sounds like a /repost tho
#
[KevinMarks]
I think your like of their share post makes sense then
#
[fluffy]
Yeah, but would it make sense to have that like on the same post that’s the same share?
#
[fluffy]
And yeah it is sort of a repost
#
aaronpk
don't override "like" for this please omg
#
[fluffy]
But like, there’s UX to consider, like how to make it appear when it’s on a page like https://beesbuzz.biz/blog/chatter/
#
[fluffy]
But what if I like their sharing it? 😛
#
[KevinMarks]
The 2 posts involved will see different meanings of their webmentions because they have different labels on the links
#
aaronpk
sure, but if i see a notification that someone liked my post, i'm not going to go click through to find out if there's anything else there
#
[fluffy]
I’m not overriding “like,” I’m saying “I like this post, and I am also sharing what was being shared”
#
[fluffy]
right and I’m not adding anything else to it, I’m just resharing the link
#
aaronpk
i thought you were also adding commentary about it?
#
[fluffy]
I mean that’s certainly a possible use case but that would be a reply or mention
#
aaronpk
so it's strictly a copy of someone else's post?
#
aaronpk
that sounds like a repost
#
[fluffy]
but their post is just a sharing of another link
#
[fluffy]
so like it might be a u-bookmark
#
aaronpk
is that bad/
#
[KevinMarks]
Right, that fits the fragmention case where you're citing part of it
#
[fluffy]
so how do I indicate that my copy of the post is both a u-bookmark and a u-repost-of?
#
aaronpk
yours isn't a bookmark, yours is a repost of a bookmark
#
[fluffy]
yes but when I repost the bookmark I want it to appear like a bookmark
#
aaronpk
appear to who?
#
aaronpk
the original post?
#
[fluffy]
appear to people looking at my notes page
#
[fluffy]
or appear to indieweb.xyz or whatever
#
[fluffy]
there’s multiple UX factors at play here. they have conflicting needs.
#
[KevinMarks]
That still works. There are 2 links in the post, so they can interpret the webmention differently
#
[fluffy]
It’s very easy to give a definitive answer to one aspect of it, but that conflicts with the definitive answer to another aspect.
#
aaronpk
i'd say if your goal is that readers should see that this post is from you, then you don't want to use repost
#
[fluffy]
I want readers to see a post that links to the original thing, but also where I got that link from.
#
aaronpk
this also gets back to what is the consuming case of this particular "via" text that you've got
#
[fluffy]
If that’s a repost of a bookmark, so be it.
#
aaronpk
if it's just human consumption you can just put that text in the "content" of the post and be done with it
#
[fluffy]
the consuming case of the “via” is notifying the other person that I reposted their link
#
[KevinMarks]
The original link can see a mention, the via link can see a like or a repost, depending on what you mean
#
[fluffy]
and I mean with all this said I have already done a thing which I think I am happy with and satisfies my usage case, and I’ve linked to it a few times in this chat already but nobody’s commenting on that particular maybe-solution
#
[fluffy]
in that my note post adds an h-cite in the content
#
[fluffy]
but h-cite feels wrong because that seems like it’s more for, like, reference/academic/whatever purposes and not “via”
#
[fluffy]
so anyway there are two UX aspects: people reading my site (I want them to see where it came from so maybe they’ll consider checking out the site I got it from) and the person running the site I got it from (I want them to see that they played a part in spreading something hopefully-nice)
#
aaronpk
i guess you can use both a "repost-of" to the "via" url and "bookmark-of" to the post that both yours link to, but that's going to take some updates to /post-type-discovery in order for people to interpret those webmentions correctly
#
[KevinMarks]
But that's always the way - you mark it up to express what you mean, and others interpret that through their lenses
#
[fluffy]
[KevinMarks] I realize that. I mean I was hoping to find out what other people have been doing and I’m just feeling like this conversation is going in endless circles.
#
aaronpk
right now post type discovery assumes a post is a single thing, you're effectively talking about a post appearing different depending on who you've sent the webmention to
#
[KevinMarks]
Which is fine
#
[fluffy]
yes, I want it to appear differently depending on who sees it
#
[fluffy]
that’s the entire point 🙂
#
aaronpk
it does sound like we don't have any existing patterns for this
#
[fluffy]
I think one of the aspects that I keep on butting my head up against is that IndieWeb stuff is sort of designed like how social networks are, rather than like how blogs are/were. Each event/note/whatever is expected to be a separate entry, but that feels silly and redundant to me when one post can be multiple things.
#
[KevinMarks]
In which Jessica Roy does sterling work in explaining the parallel meanings of a single click
#
[fluffy]
Like I find it really annoying how on Twitter and Tumblr, if someone likes and reposts a thing, I see it as two separate notifications.
#
[fluffy]
as opposed to, like, “such-and-such liked and reposted this”
#
[fluffy]
Also the “like” verb feels a bit weird in that it comes in the form of another standalone post, like it feels like “this post liked this post” and not “this person liked this post”
#
[fluffy]
but that might be my own animism encroaching on my worldview
KartikPrabhu joined the channel
#
[KevinMarks]
There is the lost nuance of star versus heart
#
[fluffy]
but then also there’s a mismatch between how indieweb makes each event a separate post and how twitter/tumblr/etc. at least give the appearance of the events being different sorts of things. A post/reblog is very different than a like; a like just appears on the notes for a post, rather than being a thing you can navigate to on its own.
#
[fluffy]
Obviously a lot of that has to do with how the data model has to work in a distributed/decentralized context.
#
[fluffy]
And I think ActivityPub also does a similar thing to h-entry in that each event is its own object that is… sorrrrrta visible at least as a JSON-LD or whatever.
#
[fluffy]
But h-entry is the only one I know of where you can actually browse to the like/bookmark/etc itself as a thing that stands on its own with some semblance of visible UX.
KartikPrabhu joined the channel
#
[fluffy]
anyway star-vs-heart is exactly why I make bookmark use ⭐️ and like use ❤ in my templates 🙂
#
[fluffy]
and RSVPs are 📅 with a subscript of ✅/❌/💭 for yes/no/maybe (“I’m thinking about it”)
#
[KevinMarks]
I like it
#
[KevinMarks]
I had a thing with my colleagues where I made a quick histogram using emoji, but they're on some retro windows version so they get a low fidelity single color version
#
[fluffy]
I feel like emoji is simultaneously the best and worst thing that ever happened to Unicode.
#
[fluffy]
and text in general.
TGiske, miklb, [jgmac1106], cbzen and [snarfed] joined the channel
#
Loqi
[[snarfed]] so yes, this sounds like a reasonable new consuming use case. it'll take a while to get wm receivers to support it, but sure, go for it
Giske, petermolnar and deathrow1 joined the channel
#
[jgmac1106]
Just remember everyone likes bookmarks but nobody bookmarks likes
petermolnar joined the channel
#
[jgmac1106]
[fluffy] at the same time I have the hardest time trying to navigate through tumblr threads to get to canonical posts or to figure out who did they actual first thing
#
[fluffy]
yeah that’s an entirely different Tumblr UX failure unfortunately
#
[fluffy]
Most templtaes do put a “originally by such-and-such” at the very bottom (for some reason)
#
[jgmac1106]
ActivityPub, form my outsider perspective seems to follow the basic logic of English grammar subject, verb, object..without realizing that rule only works...never
#
[jgmac1106]
and always...okay I am being to chatty for dev..time to sign off
[snarfed] joined the channel
#
aaronpk
[jgmac1106]: omg 😆
#
aaronpk
"everyone likes bookmarks but nobody bookmarks likes"
jacky, JKing, Freiner, gRegorLove, vika_nezrimaya, nickodd and [jeremycherfas] joined the channel
#
[jeremycherfas]
Is it actually defunct? As in no longer operational. I mean, you’re using it and I think I’ve seen it elsewhere too. My own capturing has taken a turn for the worse lately and I need to understand why.
chrisaldrich joined the channel
#
chrisaldrich
I suppose I was talking about their website and their javascript bookmarklet. The concept of what they were trying to do is obviously alive and well.
[chrisaldrich] joined the channel
#
[chrisaldrich]
I've been thinking more about [fluffy]'s question and specific example. Perhaps we could use the prior art of /person-tag which uses the combination of existing microformats `h-card` and `u-category` .
#
[chrisaldrich]
I might suggest could use the pre-existing microformats `h-cite` in combination with a `u-repost-of` , but that's already a pre-existing format for the general idea of reply. Instead, maybe we could use the combination of `<cite>` and `u-repost-of` or something similar?
#
[jeremycherfas]
For me, [fluffy] ‘s point about every post raises something interesting that I definitely share with respect to bookmarks. If I only bookmark something, I am adding it to my collection of “raw data”. If I write something myself about it or extract a quote it becomes something else that ought to include the bookmark.
#
[jeremycherfas]
Every post being just one kind of thing, I meant.
[fluffy] joined the channel
#
[fluffy]
I wasn’t expecting my random question to turn into a full-on ontological discussion
#
[jeremycherfas]
Why ever not? It is important.
#
[fluffy]
I’m not saying I’m displeased with this 🙂
#
[chrisaldrich]
s/reply/repost
Giske, [Michael_Beckwit and gRegorLove joined the channel
#
[chrisaldrich]
I've noticed that one of the originators of the Curator's Code uses a class="via" on a HT in this example: https://www.brainpickings.org/2011/10/03/charles-eames-on-design-1972/
#
Loqi
Charles Eames on Design: Rare and Wonderful Q&A from 1972
#
[chrisaldrich]
This seems an odd pattern given their clear definitions of via and hat tip.
#
[chrisaldrich]
But it's also a solid example in the wild for a potential new microformat perhaps?
Giske joined the channel
#
@webrocker
Just bookmarked (and replied to) @adactio s post on Hydration on my blog, using both 'u-in-reply-to' and 'u-bookmark-of' classes on one <a href > pointing to his url. I wonder if this works or is even supposed to work - can one webmention be two things? #indieweb
(twitter.com/_/status/1226092968546897921)
skatkov, swentel, zumid, vika_nezrimaya, [Rose], JKing and [AlisonW] joined the channel
#
Zegnat
Well, this is frustrating, somewhere my correct SSH setup is missing so I have locked myself out of my server
vika_nez` joined the channel
#
[AlisonW]
Eek! Physical access possible?
#
Zegnat
No, it is a VPS
#
[AlisonW]
(aka btdtgtts so had to travel to colo)
#
[AlisonW]
Oh poo.
#
Zegnat
I just need to backstep through a number of computer backups and figure out when I messed it up
#
Zegnat
I am pretty sure I got the right key. Maybe it is the wrong passphrase or username
#
[AlisonW]
It's (too) easily done.
#
jeremycherfas
Some day I'd like to understand the various keys I have created by obeying instructions.
#
[AlisonW]
Accidental Caps lock?
#
Zegnat
jeremycherfas, you are not alone
#
Zegnat
Found an old SSH config file, fingers crossed!
[jgmac1106] joined the channel
#
Zegnat
Success, access restored!
#
beko
Zegnat: You know that you can store ssh keys in KeePassXC and even let em automatically load using ssh-agent, don't you?
#
Zegnat
I don’t use KeePassXC, that that definitely is interesting
#
Zegnat
I am usually fine with my .ssh directory and a clear naming scheme + config file, something just went awry
#
beko
it's _local_ :D The only cloud thing involved is my own share :D
#
Zegnat
I use 1Password specifically because I also juggle a couple of shared password vaults (work team, family, etc.) And while I have a local vault in there, it never felt like it would make sense to run two password manager
#
Zegnat
*managers
#
beko
https://fedoramagazine.org/managing-credentials-with-keepassxc/ fyi. We do so too but with Keepass. It can also autoopen various vaults seperated by tab. If a tab is closed or locked keys are automatically unloaded too
#
beko
Oh and it deals perfectly with updates in the background via cloud sync or whatever.
#
Zegnat
The problem is having to migrate everyone off of one application onto another one
#
beko
ah, wasn't making you. Just thought this may be of interest. Managing ssh keys manually is a pain at a certain point.
#
beko
Maybe your tool has a similar solution, no?
vika_nezrimaya joined the channel
#
Zegnat
I don’t know, but will definitely check it out now!
[Sadik_Shahadu] and maxwelljoslyn joined the channel
#
maxwelljoslyn
chrisaldrich++ I have finally learned that h/t is supposed to mean "hat tip", not "heard through"
#
Loqi
chrisaldrich has 1 karma in this channel over the last year (29 in all channels)
[LewisCowles] and chrisaldrich joined the channel
#
[AlisonW]
is of an age where h/t means High Tension (450V+)
DeltaWhy[m] and [Michael_Beckwit joined the channel
[Christina_Hendr, jackjamieson, Scott1, [scottgruber], [jackjamieson], [cheukting_ho], m2creates and [Marlin_Forbes] joined the channel
#
@jgmac1106
↩️ Scott Gruber is going next, first #IndieWeb event in Portland three years ago. He demos http://scottgruber.me he is using @getperch and hasn't added webmention plugin yet but is adding a "now" page and pulls in his tweets and experimental art sites (https://quickthoughts.jgregorymcverry.com/s/KvnNW)
(twitter.com/_/status/1226189947369115648)
lalilulelo and geoffo joined the channel
#
@jgmac1106
↩️ Then @jackjamieson went and showed off his Wordpress site and how he uses the microsub reader. #Indieweb (https://quickthoughts.jgregorymcverry.com/s/5hg2i)
(twitter.com/_/status/1226192455051235330)
nickodd joined the channel
#
@RubygemsN
webmention-cli (1.0.0): A command-line interface for Webmention. https://rubygems.org/gems/webmention-cli
(twitter.com/_/status/1226206399840931841)
sivy, [KevinMarks] and nickodd joined the channel
#
GWG
How would you indicate a feed is a tag archive?
#
Zegnat
I would use its name
#
GWG
I meant in markup
#
GWG
rel="tag feed" ?
#
Zegnat
Aah, hmm.
#
GWG
Well, "alternate home" is supposed to be the homepage URL
#
Zegnat
Interesting
#
GWG
Zegnat: I thought so. What other combos might there be?
Nuve, [Aaron_Klemm] and maxwelljoslyn joined the channel
#
sivy
hulllooo
#
sivy
so, part of my new blog software is an indexer to move posts in markdown into a sqlite db
#
sivy
but this week I was messing around and nuked about 15 posts from the git repo
#
sivy
yikes
#
sivy
so now I just wrote "persister" to take the db contents and write it back out the filesystem :)
#
beko
swentel: Whatever you did, 0.56 fixed my account issue 🎉
#
Zegnat
sivy: nice to be able to support both directions!
#
swentel
beko, woot
#
Loqi
😄
#
swentel
good to know for future apps if I need to access accounts
#
beko
It requested a bunch of new permissions and my firstborn (no refunds!) and now it's fine :D
[chrisaldrich] joined the channel
#
swentel
beko, oh, like account/contact permissions?
#
beko
I guess it was the "identification" thingy.
#
swentel
could be, anyway, glad that one is out of the way :)
#
swentel
expecting bugs now from the photo editor ;)
#
beko
gonna disappoint you on that swentel. Snapseed is my to-go solution for this. Anything else is passed to Darktable later.
#
swentel
oh that's fine
#
beko
…and piped through mogrify -strip :D
#
swentel
wow darktable looks nice
#
swentel
still messing around with gimp
#
swentel
although I don't do /that/ much with design/photo's anyway
#
beko
me neither :D It's like a truck when I need a bicycle :D
[fluffy] joined the channel
#
swentel
wow, trying it out now, way to many options haha
#
beko
I found out how to assemble favourite options hiding the rest on default :D
#
beko
There's also this new Photoflare kid in town: http://photoflare.io/ - first contact wasn't that great tho.
#
swentel
snapseed is pretty crazy too, didn't know that app
#
swentel
good to know if people request more features, I can point them to that one, or others
#
beko
It's now Google - mixed feelings on that.
#
swentel
it's hard to get rid of them .. almost completely freed myself, need to migrate contacts/calendar to my nextcloud, then I can close my account :)
#
aaronpk
contacts/calendar is the easy part!
#
beko
Heh. Never fell in that trap :D
#
aaronpk
well as long as you don't need calendar invitation support, then it gets tricky with nextcloud again
#
swentel
hmm, no just need to share a calendar with my partner, that's all
#
beko
I used Zarafa for years. Even got a shiny certification somewhere. Zarafa is dead. Long live Kopano :)
#
[fluffy]
moving discussion on non-chronological/non-oldest-first feeds here from #indieweb-meta
#
[fluffy]
so yeah, like, reading order is definitely a thing for some things, and being able to support that in the context of a reader gets really difficult, especially if you’re trying to mingle stuff between different feeds.
#
[fluffy]
I always set my reader to do oldest-first and only show unreads, because a lot of the content I subscribe to is intended to be consumed oldest-first, and I don’t know of any content that actually is intended to be completely consumed AND read newest-first.
#
[fluffy]
newest-first always struck me as only useful for in-the-moment ephemeral stuff.
#
maxwelljoslyn
w/ newest-first anything that i actualy want to catch up on, i ineviteably end up scrolling thru its archive (if available) to find point where i last left off
#
[fluffy]
I have thought about ways of making a reading interface that allows a mixture of content, though, like ordering things based on “when it should be read” where the sort key is publishTime ASC for oldest-first, and now-publishTime ASC for newest-first
#
[fluffy]
yeah and that works fine for things where it doesn’t matter what the reading order is, but for things like serial webcomics where major plot points happen and affect later things, that can make the experience way worse.
#
[fluffy]
like, seeing the big reveal to a subplot before you even know it’s a subplot.
#
maxwelljoslyn
agreed RE newest-first useful for ephemeral. eg. I probably don't care about all the tweets someone sent while i was on vacation and will just resume checking their feed when i return (in the hypothetical world where i use twitter)
#
[fluffy]
yeah and like, most blogging stuff is very much newest-first/ephemeral. The canonical example of that for me is LiveJournal.
#
maxwelljoslyn
yeah, it's no fun to land on "Big Climax 3/3" and have the emotional impact spiled
#
[fluffy]
LJ was of course a lot of peoples’ first exposure to blogging, and was definitely where I first experienced a newest-first, no-unread-tracking timeline.
#
maxwelljoslyn
I think one thing to do in this space would be to make some test h-feeds using HTML ordered list to permalinks, with listing order NOT same as date order, and see how indie readeres display it when slurped up
#
[jgmac1106]
Scifi space operas need some rules on ordering books...
#
[jgmac1106]
Conversation today made me think back to Martys distinction between archive and feed
#
[jgmac1106]
Maybe the archive is where you dont worry about publication date
#
[fluffy]
the conceptual problem I run into with the mixed-ordering feed is that, like, how do you make it efficient? I guess implementation-wise you could have it as two database cursors, one perusing oldest-first content and one perusing newest-first, and it does a merge at presentation time.
#
[fluffy]
so okay never mind I just solved that problem, phew
#
maxwelljoslyn
ty for mentioning unread-tracking. I think that's the right answer for serial content where read order is crucial, like comics/stories - i imagine that's typically done with cookies, or user DB automatically tracks it or something)
#
maxwelljoslyn
ty for mentioning unread-tracking. I think that's the right answer for serial content where read order is crucial, like comics/stories - i imagine that's typically done with cookies, or user DB automatically tracks it or something)
#
maxwelljoslyn
ty for mentioning unread-tracking. I think that's the right answer for serial content where read order is crucial, like comics/stories - i imagine that's typically done with cookies, or user DB automatically tracks it or something)
#
maxwelljoslyn
gah! sorry y'all
#
beko
happens :)
[snarfed] joined the channel
#
[snarfed]
^ this is jameysharp's pet issue, he advocates for robust paging and RFC5005 support in everything. eg https://github.com/snarfed/granary/issues/152 . he's not wrong!
#
Loqi
[jameysharp] #152 Atom/RSS feed paging
#
maxwelljoslyn
snarfed++ thank you! [chrisaldrich] mentioned a relevant RFC but didn't recall the number
#
Loqi
snarfed has 51 karma in this channel over the last year (88 in all channels)
#
maxwelljoslyn
what is next?
#
Loqi
It looks like we don't have a page for "next" yet. Would you like to create it? (Or just say "next is ____", a sentence describing the term)
#
maxwelljoslyn
what is rel-next?
#
Loqi
It looks like we don't have a page for "rel-next" yet. Would you like to create it? (Or just say "rel-next is ____", a sentence describing the term)
#
[fluffy]
yeah Jamey got me on board with 5005 and I made a big point of supporting that in Publ
#
[fluffy]
so all of my website templates support it
#
[fluffy]
well, all of my atom feeds, obviously
#
[fluffy]
I think I also have robust rel=prev/next on all of my index templates
#
[fluffy]
but there’s probably some places where that’s slipped
#
maxwelljoslyn
what's the diff between paging and rel-prev/next? is rel-prev/next about moving from one post to its successor/predecessor, while paging is about loading only some number of things at a time?
#
[fluffy]
yeah, pagination is the concept of having pages of things, prev/next is the relationship between those pages.
#
[fluffy]
RFC5005 is fundamentally an explanation of both things together.
#
[fluffy]
Most RSS/Atom feeds only display the most recent page of content, with no ability to use the feed to traverse the entire archive. 5005 specifies how to make that doable.
#
maxwelljoslyn
ty. I am not at the point where I have enough stuff loading on one page (eg blogfeed) to use prev-next for that, am only using it for links between posts atm
[LewisCowles] joined the channel
#
maxwelljoslyn
Understood
#
[fluffy]
A big part of the context for why Jamey got into this is his site Comic Rocket tries to be a comics-oriented reader, and most comics’ feeds are useless for the archival purpose so he had to write a rather fragile scraper to try to find the archive via common patterns that a lot of comics CMSes use.
[Rose] joined the channel
#
[fluffy]
Comic Rocket also provides an “archive binge” function where someone can build their own rerun feed that provides comics from oldest to newest with a specified number of updates per day. With 5005, such a thing isn’t necessary, because feed readers can support that themselves.
KartikPrabhu joined the channel
#
maxwelljoslyn
getting all that info into the feed as metdata, making hte feed more self-contained
#
maxwelljoslyn
oh! fluffy! IIRC you have strong AWS knowledge - would you be willing to spend a little time tomorrow helping me figure out why my S3 redirect(s) aren't working? I swear I've followed the docs to the letter but neither method (per-file or routing rules) has worked and it's been frustrating me for dayz
[jeremycherfas], TrondGiske, chrisaldrich, [Michael_Beckwit, KartikPrabhu and [snarfed] joined the channel