#dev 2026-10-07
2026-10-07 UTC
cotsuka, ghorsey_mobile and somannak joined the channel
# 
carrvo Am I misunderstanding something? To use IndieAuth without introspection does the client actually need to do anything different? Or is that entirely on the service/idp to do things?
grufwub joined the channel
# carrvo Sorry, wrong terminology. It is ok for the browser to behave the same way regardless of introspection? And the client chooses whether to query the idp?
# carrvo I found that my IndieAuth bot assumes that a service will use introspection and I am trying to figure out what I need to change to support simpler scenarios.
# carrvo The bot does browser, the profile, and the idp.
# carrvo Normally if I were to log into the indiebookclub: I use a browser, then indiebookclub queries my profile, then my idp, then redirects me to my idp, which redirects me back to indiebookclub. The bot does everything that indiebookclub does not.
# carrvo But does it need to behave differently depending on whether indiebookclub uses introspection? The idp portion does...anything else?
# carrvo I'm going to try my bot against indiebookclub...I had been doing it against IndieAuth.com
# carrvo No, it is not required in the spec. That is why I want it optional for my bot. (I didn't for my site.)
# carrvo Hmm...invalid IDP: the scheme is invalid.
# carrvo I think that is an issue on my end...
# carrvo A different one.
# carrvo gRegor++ at least now I have a second service to test against.
cotsuka and ghorsey_mobile joined the channel
# 
gRegor There's the relatively new https://indieauth.rocks/ too
ghorsey_mobile, Guest5158 and ben joined the channel
# carrvo Looks like indiebookclub does not support user paths (only domains).
# carrvo Looks like IndieAuth.com does not include its client_id in the redirect from the log-in endpoint. But I do when requesting it.
# carrvo IndieAuth.rocks, if I recall from recently testing my main implementation, wants user interaction I have not programmed into my bot. I tried it anyway and...I think some test run started? But it is not like I will be able to access the results.
# carrvo aaronpk: my logs show a client id of f59w6zsfkzm6jyj0 and a run id of 01M4A8KP39F7CS6RFPRKZP0SFG if it is of interest.
# carrvo The comment about the logs was for IndieAuth.rocks
# carrvo Also, there is no way my bot can support it's replacement so it spurred me to test against an older client before it disappeared.
# carrvo Any other suggestions for a place I can see whether my bot can simply login?
# carrvo As a test.
# carrvo Hey tantek! On a different note I sent you a private message a few days ago unrelated to the IndieWeb.
# carrvo About to try your suggestion.
ghorsey_mobile joined the channel
# carrvo Sorry, Discord. Not important in the least.
# carrvo I don't see any error with IndieLogin.com! thanks tantek++
# carrvo Or, I get code_expired when I dig deeper. Hmm...something to keep working at.
srazkvt and [pfefferle] joined the channel
# 
WinnerWind webmention question: is it bad form to re-send webmentions on site build even if nothing has changed?
# offeringofpie I don't think it counts as double, so it should be fine š
# kensanata [artlung]: Yeah, that issue 63 for human.json was me.
# kensanata [artlung]: You blog post made me look for my own FOAF blogging and I found a page from 2004 on how to locate friends on a map using xplanet. And at the top it says: "(Looking back at all this about twenty years later: what a privacy nightmare!!)"
# kensanata WinnerWind: Back when I had a webmention implementation, those would add comments to my blog and I ended up switching it off because I was looking for a signal ("somebody is linking to this page and has something to say, I should read it") and what I was getting was a kind of noise: rebuilds, off-hand mentions with links on pages with unrelated
# kensanata topics, etc.
# WinnerWind kensanata: that's kind of the nature of webmentions, off hand mentions tend to be kind of common and i guess that's also the point of webmentions
# 
Loqi It looks like we don't have a page for "offhand mention" yet. Would you like to create it? (Or just say "offhand mention is ____", a sentence describing the term)
# WinnerWind [tantek]: i assume they meant "this person linked to me but for no particular reason"
# kensanata WinnerWind: I don't think there's an inherent definition of what webmentions are for. All I can say that the way people used them was not what I wanted from them.
# WinnerWind but im more concerned about rebuilds because at the moment my sending script just resends every webmention (takes about 1 minute at the current size of ā100 pages) to every site it can find which supports webmentions, like telegraph but bulk
# WinnerWind driveby mention?
# kensanata Offhand mentions are things like a very long blogpost on some topic I am not interested in, with one or two sentences related to something I blogged about, and a link to my site. So when I follow back to the blog post, there are 998 sentences I don't care about.
# offeringofpie SEO farms
# kensanata Which is unlike a comment on my blog. So webmentions ended up not being a decentralised way of commenting on my blog.
# kensanata Which is what I wanted.
# WinnerWind [tantek]: yup
# WinnerWind i just hope that re-sending webmentions on every site generation (which happens every time i add new content/ fix a bug or something) isn't too critical for webmentions, hoping that making minor changes like fiixng the h-card and stuff doesnt break things either
# WinnerWind https://indieweb.org/spam#Webmention oh cool there's a page on this
Saige_Leah, dduckling, jjuran and hypha_ joined the channel
# 
simpleauthority Well, the spec does call to do that right? if you update something, you should re-send the WM so target can check what changed.
# kensanata back to Alice's original post. Bob's publishing software sends a Webmention to Alice notifying that her article was replied to, and Alice's software can show that reply as a comment on the original post."
# kensanata I wouldn't focus on the technical aspect too much. If you start with the first paragraph of the introduction of the W3C recommendation, it starts with an example: "A Webmention is a notification that one URL links to another. For example, Alice writes an interesting post on her blog. Bob then writes a response to her post on his own site, linking
# kensanata In this example, webmentions help humans react to each other. It's an implicit invitation to interact. A rebuild of the site, a minor change to the text, or even a significant but unrelated change might not merit a webmention.
# kensanata A possible UI for webmentions would be: after editing a page, review the link targets and carefully pick the ones you want notified.
# kensanata From this perspective, webmentions are sent with intent, not automatically.
# offeringofpie I tried to automate sending webmentions to outgoing links on my blog posts, but it ended up being a long ping pong process. it's a manual process for me now
# kensanata The solution I used for a few years after dropping webmentions was referrer information (referer header). It meant that somebody actually followed a link (instead of somebody adding a link). I then checked the source for the actual link, the request for user agent and all that, in order to avoid spam, checked for duplicates, and if it all worked
# kensanata out, I counted that as a signal. Many sites no longer allow browsers to send referrer info, for privacy reasons, of course.
# offeringofpie ...I do that most times, `noreferrer nofollow` sorry
# WinnerWind what does that do @offeringofpie?
# WinnerWind kensanata: telegraph works like that, actually!
# WinnerWind you can automate it but I use telegraph to resend webmentions for testing purposes
# 
rubenwardy Yeah privacy controls have broken refback. Even without noreferrer most browsers will send the site name and not the page
# offeringofpie winnerwind[d] if you add `rel="noreferrer nofollow"` to external anchor links, you're telling the browser that you don't want to send a referrer header (which exposes which site visitors came from), and nofollow is like a declaration that you don't endorse the site the link is about. I'm blurry if nofollow does anything super practical tho.
# offeringofpie I kept it out of habit tbh
# WinnerWind i see ^_^
# WinnerWind i didnt even know why referrs work
# WinnerWind seems like an odd header to send
# offeringofpie in analytics services you can sort incoming traffic and get very useful data. It's like `?utm_source=<source>` but without polluting the url.
# 
[tantek] well-behaved* crawlers, in contrast to say /LLM_traffic
cotsuka joined the channel
# WinnerWind every crawler ive seen goes buck wild when given an <a tag
# WinnerWind ive been monitoring 5 crawlers on my site (after trapping them in a maze) and even though each page's <a> tag is technically inaccessible, they follow them anyway
# 
[KevinMarks] https links withhold referer by default anyway
# rubenwardy I think even well behaved crawlers follow these
# rubenwardy nofollow just means don't trust this link for seo reasons
[Murray], cotsuka and CRISPR joined the channel
# carrvo With Webmentions you don't necessarily have to accept them as a comment on your site. You could add some kind of check to see whether it is a comment or not. For example, some use microformats to distinguish between comments, likes, RSVPs, etc.
# carrvo That ties in nicely to a conversation recently where someone did not have meaningful interactions with microformats--they pair nicely with Webmentions.
[social], cotsuka, Hollis and ghorsey_mobile joined the channel
# WinnerWind carrvo: i'm pretty sure all receievers check whether its a comment, like, rsvp etc.
ghorsey_mobile, ax562, cotsuka, doesnm__ and barnaby joined the channel
cotsuka, ghorsey_mobile, CrimeWave, bterry, mary2, [Leo] and [aciccarello] joined the channel