#dev 2025-08-22

2025-08-22 UTC
grufwub joined the channel
#
[tantek]
What is Ghost?
#
Loqi
Ghost is an an open source CMS for blogging, general web publishing, newsletters, and paid subscriptions, and also a hosted silo at https://ghost.org https://indieweb.org/Ghost
#
[tantek]
a2adam ^ check for tips there or ping the folks listed in IndieWeb examples and ask them how they did it
CRISPR joined the channel
#
[snarfed]
I need to change that error message, it's misleading
solonovamax joined the channel
#
solonovamax
I remembering there being some kind of microformat or something using a sidecar file for listing all of the other websites you have accounts on, but I cannot for the life of me remember what it was, so I'm wondering if anyone else would know it off the top of their head.
#
solonovamax
oopsies, typo there, s/remembering/remember/
#
[snarfed]
rel=me?
#
solonovamax
I remember there being something else, but I could just be imagining it
#
solonovamax
pretty sure I remember it being json-based as well
#
.pi.r2.
maybe based on json-ld? there should be plenty of schema to link to other profiles
#
solonovamax
possibly, but I just can't remember what the format was
[morganm], CrimeWave, just_jimmy and bitauger joined the channel
#
capjamesg
[snarfed] Does Granary do any caching?
#
capjamesg
I was checking my logs and found 6k requests from granary to my RSS feed.
cow_2001, ttybitnik, duanin2, chenghiz_ and [schmarty] joined the channel
#
[tantek]
What is a side file
#
Loqi
The sidefile-antipattern is a violation of the DRY principle by the use of secondary files (typically in some one-off XML format) to provide information that is a duplicate of information available in primary files on a website (in HTML), and is an antipattern due to typical DRY violation problems such as out-of-date, missing, corrupted, or outright false data https://indieweb.org/side_file
#
[tantek]
What is a sidecar file
#
Loqi
It looks like we don't have a page for "sidecar file" yet. Would you like to create it? (Or just say "sidecar file is ____", a sentence describing the term)
#
capjamesg
Sorry, I meant h-feed.
#
[snarfed]
capjamesg yes! http://granary.io the service, right? it caches for 5m
#
[snarfed]
it doesn't do any polling, though, and it doesn't know the semantics of whatever's calling it, so I'm reluctant to do anything more aggressive or "smart"
#
[snarfed]
sorry for the load! 6k requests over what time period?
#
.pi.r2.
I don't buy the argument against "sidecare files" fully. my blog is markdown sources, JSON feed is no more a DRY violation than HTML pages for me. sometimes the community seems so passionate about HTML it appears to me opinionated in ways that have little to do with defending the "mission" laid out at https://indieweb.org as far as I can tell
#
[snarfed]
(to rephrase, http://granary.io never fetches your site on its own. it only fetches when it gets a request for your site, and even then, that should generally be at most once every 5m per URL)
#
doesnm
maybe the point is to use h-feed instead of rss/atom?
#
.pi.r2.
I don't want my landing page to contain excerpts of the articles' content, personally
#
.pi.r2.
let alone full articles
#
.pi.r2.
I have the newspaper view that would qualify as a reasonable h-feed, but see little value in it vs a dedicated feed with a more bounded history
#
[artlung]
RSS / Atom / Feed files might be an antipattern but the concept works.
#
[snarfed]
h-feed can still do excerpts. I do the same thing, excerpts only on my front page, and the h-entries have u-urls pointing to the permalinks with full content
#
.pi.r2.
right so your reader can't show the full content though
#
.pi.r2.
I feel readers should have a dedicated protocol with cursors to retrieve only what changed since last pull anyway, so it feels all antipatterny
#
[snarfed]
h-feed readers can fetch the u-url to get the full content
#
[snarfed]
the dedicated protocol evolution was toward websub, to get rid of polling altogether
#
[snarfed]
but sadly afaik that never quite hit mainstream adoption
#
[snarfed]
(or maybe it did? not sure)
#
[artlung]
What is websub?
#
Loqi
WebSub is an open standard (W3C Recommendation) notification-based protocol for web publishing and subscribing to streams and legacy feed files in real time, previously known as PubSubHubbub or PuSH, and briefly PubSub https://indieweb.org/websub
#
[artlung]
It’s documented, yup.
#
.pi.r2.
slack doesn't even support non-XML formats 🙁
#
.pi.r2.
that's my only use for feeds, to cross post to my personal channel
#
[artlung]
For a time rss and atom feeds would use the `feed://` protocol. Was part of that related to what Pierre’s question implies? It never quite worked right in all browsers.
#
.pi.r2.
oh that's just the URL scheme
#
[artlung]
Sure, but a url protocol can imply common semantics: Mailto: news:
bebethatmustbeme joined the channel
#
.pi.r2.
I mean dedicated protocol as in something that could be built on top of HTTP, eg through headers (could be ETag and Changes-Since for example)
#
[artlung]
Not saying it worked (it obviously didn’t catch on) but curious if there’s a relationship
#
.pi.r2.
that avoids the stateless feed that returns the same stuff over and over, no pagination, no memory of what's already received
#
.pi.r2.
websub could be great, I'll have a look. as usual the problem is going to be adoption
#
.pi.r2.
generally OK with polling if it can be made really cheap most of the time
#
[artlung]
Might be interesting to do a websub carnival. No idea what would justify it but I was looking at the concept a few months ago since I create and repurpose feeds in a few places.
#
.pi.r2.
I don't love push models given they require infra staying up and accessible externally and/or queues with retries
#
[artlung]
Yes, creates opportunity for spof
#
[schmarty]
the cool thing about websub is that, while it enables real-time updates through push, it has this great fallback in that polling is always available.
#
[schmarty]
the unfortunate thing about websub is that _polling is always available_
#
[snarfed]
minor point, feed pagination does exist, if (again) afaik not widely adopted, https://tools.ietf.org/html/rfc5005
#
[tantek]
feed pagination is among many features in Atom that never got consuming code adoption
#
[tantek]
It's a good example of why we need consuming code examples to validate any attempt at making new (micro)formats
#
[schmarty]
we have notes about RFC 5005 from discussions at indieweb events dating back to 2018 😅
marc_in_space, NaomiAmethyst, ttybitnik and duanin2 joined the channel
#
Loqi
[preview] [jameysharp] #152 Atom/RSS feed paging
gRegor and barnaby joined the channel
#
[artlung]
“Context escape” is a useful term. Could retroactively apply to several foofaraws and legitimate controversies.
barnaby joined the channel
#
[tantek]
happens quite often indeed. also the start of that post makes me want to catalog a "top 10 github flameout issues" for the lulz
#
[tantek]
the JS ":" or not in the scheme controversy (node, natch) is one of my faves
#
[artlung]
“How communication fails” … lulz, sure, but the importance of setting a table for a rich communication. Cautionary tales.
#
[mattl]
> “top 10 github flameout issues”
#
[mattl]
And who said late night TV was dead?
#
[schmarty]
sounds like the old drama wikis 😬
memooo joined the channel
#
capjamesg
[mattl] Too soon. Colbert 😭
#
[mattl]
I wish I liked any of those shows… I tried to watch them when I first moved here but never got into it.
#
[tantek]
Colbert 🫗
#
[tantek]
how is there not a weekly podcast call "Main Characters" that summarizes the main characters from the week 😂
#
[mattl]
there’s probably a reddit thing for it, but meh
CrimeWave and btrem joined the channel
#
[mattl]
Picked up a cheap little computer for a home server. https://www.ebay.com/itm/357459567933 — it just arrived. It’s brand new, which is very surprising.
opencircuit and ttybitnik joined the channel