#dev 2025-03-21

2025-03-21 UTC
sebbu2 and glacier joined the channel
Lux joined the channel
#
trwnh
immibis re: "json with an extra context tacked on", yes that is essentially AS2 assuming you don't use any extensions to the context. the problem is a) everyone uses extensions, b) semantic divergence even within the normative terms
#
trwnh
it's a problem if you expect people to share your understanding when you actually have different worldviews
#
trwnh
it's like the example with h-cite vs u-repost-of i brought up yesterday
#
trwnh
people need to agree on *protocol* and *semantics* not just data format. this is why concept modeling is important
grufwub, skbohra and Xe joined the channel
#
[tantek]
Does anyone remember that recent post critical of Bluesky that said things like all posts are permanently public and other things? Can't seem to find it
pfych, Guest6, joshl and ainstein joined the channel
#
ainstein
Hello! Is the discord server down? Trying to join shows that there are no text channels in the server.
bbbhltz, Guest6_ and Guest6 joined the channel
nemonical and [Murray] joined the channel
#
pfych
On standardized alt text - It does turn out that there's a standard image description tag: "Exif.Image.ImageDescription" but it's limited to ASCII, "Exif.Photo.UserComment" exists as well & isn't limited. I'm surprised more photo apps don't expose one of them as alt text!
#
Loqi
[preview] [enmaku] #1266 Set initial alt text from exif data if available
#
[Murray]
There have been a couple of pushes to do so, I believe. Flickr pulls out the comment as a caption on upload. One issue is that sites will often strip EXIF data on upload, to prevent people doxxing themselves, amongst other things
#
[Murray]
but yeah, it does feel like this should be standardised somehow
#
pfych
I feel like there definitely needs to be a push from some large body to have a dedicated "alt" field that everyone agrees not to put personal information in and sites wont strip haha
#
[Murray]
one can dream πŸ˜„
#
pfych
There's just a lot of places it could go which I think makes it hard to pick one to be the standard... (XMP: XMP-iptcCore:AltTextAccessibility, XMP: XMP-iptcCore:ExtDescrAccessibility, Exif.Photo.UserComment, etc.)
#
pfych
Hopefully something comes around soon!
#
pfych
For context - I've got about 400+ photos on a gallery page on my website which completely lack alt text right now and It'd be great to have that done in some standard way. For now maybe a meta file might be the way to go, since the page is built programmatically ... https://pfy.ch/images.html
Guest6, [baudouin], gregsarjeant, sivoais and [Laptop_Philhar] joined the channel
#
capjamesg[d]
I have been wanting to write this for a few weeks now: I have been wanting to write this for a few
#
jeremycherfas
capjamesg[d]++ for a useful post that I will try and internalise. I usually say feed reader rather than RSS reader, but you are right, web reader is a more approachable description.
#
Loqi
capjamesg[d] has 48 karma in this channel over the last year (214 in all channels)
#
[qubyte]
Hmm. Feed reader is already common parlance though, isn’t it?
#
capjamesg[d]
I have separate reservations with feed readers πŸ˜„
#
capjamesg[d]
People, not feeds.
#
capjamesg[d]
One nice thing about Mastodon, etc. is that you press one button and you see what someone has written. But with web readers you have to find the feed URL, choose which one you want, confirm it's the one you want, then you'll start seeing the content.
#
[qubyte]
πŸ˜• it’s not reading a person though. It’s reading a feed of things. A feed might be nothing to do with a person or even a web site.
#
jeremycherfas
Which is, I think, why web reader is actually a better term.
#
jeremycherfas
I mean, I agree with capjamesg[d] that it is a better term.
#
[qubyte]
It seems too general to me. To the point of no longer being descriptive. I’m just coming out of a long meeting though, so my brain might not be working right. πŸ˜…
#
jeremycherfas
You're excused. :)
#
[baudouin]
just landed - would this be the rigth place to discuss A11
#
[baudouin]
A11Y coding and best practices ?
shadfi joined the channel
#
capjamesg[d]
Yep! This is the channel for discussing web programming.
GuestZero and Kolev joined the channel
#
Kolev
pfych, I am lazy and don't always add alt text.
nemonical and GuestZero joined the channel
#
[tantek]
capjamesg[d]++ great post. Webreader++
#
Loqi
Webreader has 1 karma over the last year
#
Loqi
capjamesg[d] has 49 karma in this channel over the last year (215 in all channels)
#
[tantek]
RSSreader--
#
Loqi
RSSreader has -1 karma over the last year
nemonical, nemonical1 and b0rg joined the channel
#
Lars-Christian
I like what you're saying there @capjamesg[d]
#
Lars-Christian
I also usually say "feed reader" but I think "web reader" is a much better term
#
[Murray]
(Craft is tangential, I think this should work with Twig/PHP in general)
GuestZero_ joined the channel
#
[artlung]1
Has anyone seen blog comments as GitHub (or other source control) pull requests | branches? Like literally each comment would be a pull request - maybe a single file named with a combined key of `post-slug` + `YYYY-MM-DD-HH:MM::SS` file? If one provided an issue issue templates it might be weirdly straightforward. It is absolutely the wrong way to use Github but I'm sure that's just me being uptight and it seems utterly plausible.
#
carrvo
I mean, I find it strange that GIT is used for blog source (pre-HTML) in the first place.
Luther_Gray and Pixi` joined the channel
#
[aciccarello]
There are some tools for comments from github issues, but not necessarily PRs
#
[aciccarello]
What is github?
#
Loqi
GitHub is a specialized content hosting silo for code, issues, comments, and static content that has some aspects of a commons, and through free domain mapping, a content hosting service as well https://indieweb.org/Github
#
[aciccarello]
https://utteranc.es/ was the one I was thinking of but there are more on that page
#
[aciccarello]
It would be an interesting way to review comments though
#
[artlung]1
[aciccarello]++ for remembering a clever adjacent idea. I've not seen http://utteranc.es before. _"The street finds its own uses for things."_
#
Loqi
[aciccarello] has 7 karma in this channel over the last year (31 in all channels)
jjuran, nemonical, [mattl] and [snarfed] joined the channel
#
[snarfed]
it seems awkward at best
#
[aciccarello]
I see the appeal for developer blogs but I wouldn't want to limit comments on my site to GitHub users
#
[artlung]1
I've been revisiting how my sites work and where their "sources of truth" are. Comments in a database for me, or in an external service. And it occurred to me if I migrated visitor comments to the filesystem and each was a file then with the right public architecture I could cut out the middleman.
[schmarty] joined the channel
#
[artlung]1
πŸ’― yes, that's the ultimate downside. If 100% of my new users had GitHub accounts a case could be made. But, no.
balintm, btrem, jjuran and Pixi joined the channel
#
btrem
I'm thinking of adding oauth/openid to a Drupal site for a housing coop's website. But I really don't understand how it works. I'm trying to wrap my head around it. IIUC, oauth works if a user creates an account using an id provider. But can they use it if they already have an account?
#
btrem
I ask because the website in question is not open to new members. It is only for residents/coop members, and I create the account. So is this something I can do? It'd save me from the endless "I forgot my password queries."
#
gRegor
I don't have direct experience with this, but it should be possible to sign in to the local Drupal account, then go through some flow of "link OAuth" (external identity provider). Once they successfully OAuth and they're linked, they could log in by either local Drupal or OAuth.
#
gRegor
Depends on if the Drupal plugin supports it
#
btrem
I was thinking that I can't really test it on a docker dev site because I don't have social media accounts, but I just noticed that the plugin supports GitHub. That I do have. So I think I'll install it and see how it works.
ttybitnik and [Ana_R] joined the channel
#
[Ana_R]
Hi friends - has anyone have recent examples of an "updated" note/aside on your posts? https://indieweb.org/updated
#
[Ana_R]
I have a microformat question which is dev too but I will post on the microformats channel.
#
pfych
Ana_R - I have them at the bottom of the posts. See: https://pfy.ch/music/scrobbler.html I think it looks less cluttered. I do need to look into the dt-updated tag though since I don’t use that!
#
pfych
On my homepage as well I use those same dates to render sparkle emojis next to posts updated within the last 30 days, with text on desktop.
[jeremycherfas] joined the channel
#
gRegor
I commented over there, but it's fine to ask mf2 questions like that here too since it's about indieweb posts :)
#
gRegor
Here's one of my examples with an updated date: https://gregorlove.com/2025/01/new-year-new-personal/
#
gRegor
what is updated?
#
Loqi
An updated is a post that has been edited by the author https://indieweb.org/updated
#
Loqi
[preview] [gRegor Morrill] New year, new personal website? (Or update your existing site?) Join us for a virtual IndieWeb meetup, Homebrew Website Club, January 8th at 6PM Pacific!
#
capjamesg[d]
Have there been any discussions on permalink design for web resources pertaining to locations?
#
capjamesg[d]
Breakfast and Coffee didn't have a structured permalink design.
#
capjamesg[d]
I wondered if there was a best practice for this sort of thing like /country/city/<location-name>
#
gRegor
Wonder what foursquare does/did
#
gRegor
what is foursquare
#
Loqi
Foursquare is a venue reviews, recommendations, and lists silo based on its previous location-based checkin support, a feature that was spunout to the separate Swarm application which still shares some data back & forth with Foursquare https://indieweb.org/Foursquare
#
[aciccarello]
Locations are hard because they are squishy. City boundaries can change. Restaurants move. Chains are spread across multiple locations.
#
gRegor
My first thought is if you have sub paths like that, ideally each of those sub-paths would return a list of matching venues. Then my next thought is whether that's actually useful, e.g. a country level list could be unwieldy.
#
gRegor
If it's a wiki, like I think it was before, categories probably work better
#
capjamesg[d]
I think /uk/ would return a list of all cities. /uk/edinburgh would return Edinburgh coffee shops. /uk/edinburgh/<location> would return a specific coffee shop.
#
gRegor
categories/tagging could be a lot more flexible with those squishy boundaries
#
[aciccarello]
I agree with gRegor, it's easier to have category pages for that structure
#
capjamesg[d]
This is an interesting design problem.
#
gRegor
good point, list of regions/cities makes sense at the country level
#
capjamesg[d]
I used categories for this in the last iteration.
#
gRegor
Is it going to be MediaWiki again or tbd?
#
capjamesg[d]
But the URL UX per se was poor.
#
capjamesg[d]
gRegor I am giving it some thought.
#
capjamesg[d]
I'm not sure if I want to start this yet, but I do want to think about it.
#
gRegor
You can embed some pretty neat category trees with MW, we don't use it a lot but here's one example: https://indieweb.org/pop-up#Past
#
capjamesg[d]
Is #categorytree native?
#
capjamesg[d]
That's cool!
#
gRegor
I mean I think it's in MW core, probably on by default? May just need a config
#
capjamesg[d]
I love MW but running a wiki involves a lot of admintax.
#
gRegor
Some of the year categories use it too https://indieweb.org/Category:2024
#
gRegor
Definitely, especially anti-spam unless you really lock down login.
#
gRegor
what is dokuwiki
#
Loqi
DokuWiki is an open-source wiki-project that uses file-storage https://indieweb.org/DokuWiki
#
gRegor
Another option^
#
capjamesg[d]
The alternative idea in my head is a news aggregator for all things breakfast and coffee πŸ˜„
#
gRegor
ooh, nice
gRegorLove_ joined the channel
#
btrem
capjamesg[d]: I don't know what "Breakfast and Coffee" is. So with that caveat, I'd probably use /country/city only for feeds, that is, for a list of coffee shops in a city or country or whatever. But I'd use a more durable url scheme for any individual post. For me, that'd be /yyyy/mm-foo, e.g., 2025-03-java-shop.
#
capjamesg[d]
What is breakfast and coffee?
#
Loqi
Breakfast and Coffee was a wiki to which one can syndicate reviews of and information about breakfast and coffee stops published on their own website https://indieweb.org/Breakfast_and_Coffee
#
capjamesg[d]
It is no longer active, but that was the idea ^
#
[tantek]
capjamesg[d] pretty sure I documented prior art for locations in URL segments somewhere
#
[tantek]
what is Dopplr
#
Loqi
Dopplr was a silo (at dopplr.com) for sharing future travel plans https://indieweb.org/Dopplr
#
[tantek]
not saying that's the "best" or "ideal" but at least one previously implemented URL design
#
btrem
No time to read the indieweb page (but somehow I have time to chat here!), but I get the gist correctly, then I'd definitely use some non-hierarchical url for individual reviews, and use location urls as feeds, sort of like tags on many a site. That's my 2 pence.
gRegorLove__ and gRegorLove_ joined the channel
#
[tantek]
I think location URLs would well as maps as well, possibly aggregating icons for permalinks to posts that fit within the location polygon they represent
#
[tantek]
would work* as maps as well
gRegor joined the channel
#
[tantek]
I have figured out a location shortURL design that I like but not a location URL design, yet, especially one that uses named places (which I understand the appeal of)
gRegorLove_, geauxweisbeck4 and balintm joined the channel