#dev 2021-09-29

2021-09-29 UTC
Seirdy, superkuh, [manton], [tw2113_Slack_], KartikPrabhu, marksuth and [jacky]1 joined the channel
#
[tantek]
GWG, [manton], I stubbed a Discovery section for h-feed, take a look and LMK what you think! https://microformats.org/wiki/h-feed#Discovery (does it make sense, is it implementable, what stronger/clearer guidance would you look for as an implementer?)
#
GWG
[tantek]: It looks good so far. Few questions
#
GWG
Let's say the main page of my site, where my h-card is, is not an h-feed. How would I indicate that there is a feed elsewhere on that page where people are looking? What if I have multiple feeds I want to show?
#
GWG
For example, let's say I'm tantek.com and have an updates and an events feed?
#
GWG
Re the mf2 parsing specification and implied h-feed...
#
GWG
Zegnat had some interesting thoughts on that.
#
[tantek]
GWG, view the source on tantek.com to see how you can do link rel=alternate discovery of my updates feed. I could add another link rel element for the events and articles feeds as well.
#
[tantek]
GWG, yeah the IndieWeb h-feed page probably needs some updating, most of the thoughts/brainstorms there are out of date compared to current discussions we've had
#
GWG
They are out of date. I think some of the ideas could be incorporated.
#
[tantek]
maybe. some of them need discussion and likely discarding as well.
#
GWG
Agreed.
#
[tantek]
like I'm pretty sure I convinced snarfed that rel=feed is a bad idea for h-feed discovery
#
[tantek]
(it makes it harder (more work) for "traditional" feed readers to discover h-feeds compared to feed files)
#
[tantek]
maybe it's worth documenting rejected ideas?
#
GWG
But there are a few ideas there I sort of like for the spec
#
GWG
For example, the section on multiple h-feeds on the same page.
#
[tantek]
the new Discovery section covers multiple h-feeds on the same page. what's missing?
#
GWG
"If the publisher wants secondary h-feeds to be referenceable , they must put unique ID attributes on them"
#
GWG
That's consistent with the parsing of ids we previously introduced.
#
GWG
And "Thus if consuming code wants to reference a secondary h-feed on a page it must do so with a url with the fragment ID of that secondary h-feed"
#
[tantek]
ah, publishing advice, yes
#
GWG
Back to the comment of an implementation only needing one h-feed, what about the scenario where there is more than text/mf2+html link?
#
GWG
I'm thinking of WordPress, where by default, a site has a main and a comments RSS feed.
#
[tantek]
right, same question in both cases
#
GWG
I may be overthinking this.
#
GWG
But I think about the firehose idea.
#
GWG
Where some people have a suggested feed and an all feed.
#
GWG
Where one is what they think you want, and the other is everything
#
GWG
But, circling back to your Discovery section, it is good.
#
[tantek]
yes the question of how to publish different subsets, and how to consume them, is still a difficult UX question, nevermind the plumbing
#
GWG
In implied h-feed...you may need to consider the way some people structure their page. They have a top-level h-card, and everything else is an h-entry.
#
[tantek]
I still believe in picking a smart default (of post "types" - implied), and then making it VERY easy for users to "mute" specific "types" of posts from specific people
#
[tantek]
rather than trying to come up with a zillion feed variants to publish
#
[tantek]
that's too much work for publishers
#
GWG
I'm just thinking about what people do.
#
[tantek]
ideally there's only a requirement to have *one* feed on your site
#
[tantek]
I'm thinking about what's the least work we can ask people to do to make the system "work"
#
GWG
Accepted
#
[tantek]
there's always going to be edgecases that publish numerous feeds
#
[tantek]
that's not the point of a standard (edgecases), the point of a standard is to optimize for the common case
#
[tantek]
and spend *most* of your time thinking about the common cases
#
[tantek]
if you're spending most of your time thinking about edgecases, you're misprioritizing your work
#
[tantek]
(in terms of specs)
#
GWG
Moving to the statement under implied feeds...are we adopting the idea into the spec that h-feeds do not have to be just h-entrys?
#
GWG
It's a proposal I think we have enough examples of in the wild to amend the spec.
#
[tantek]
indeed I think the idea has merit
#
[tantek]
and there doesn't seem to be any opposition
#
[tantek]
if you want, you can link up the issue into that spot in the spec
#
GWG
The other question is, should an h-feed be all of the same type?
#
GWG
Or a mix of types?
#
GWG
I don't see why it wouldn't be a mix
#
[tantek]
mixes can be allowed IMO
#
GWG
So, there is a common pattern that I am thinking of. No h-feed, but a page of h-entrys and a singular h-card.
#
GWG
I don't think it is an edge case.
#
GWG
I'm jumping around a bit.
#
GWG
I'm looking at the Use Cases section of the h-feed spec page. I think I can expand that
#
GWG
That might help in discovery
#
[tantek]
if you're going to expand that, it would help to cite actual examples in the wild
#
[tantek]
rather than theoretical "what if one h-card and many h-entrys"
#
GWG
[tantek]: aaronpk's site has the one h-card and the h-entrys, for example.
#
[tantek]
because my immediate answer is, if the publisher wants to do that, they can wrap the whole thing in an h-feed with a u-author on the one h-card
#
[tantek]
aaronpk can definitely do that 😛
#
GWG
[tantek]: That's why I was looking.
#
[tantek]
no need to do any implied logic
#
[tantek]
this is what I mean by not designing a spec around edgecases
#
GWG
When I went into /h-feed and updated all the indieweb examples, I was looking about what people were doing with h-feed properties like author and name.
#
GWG
That was 2019-07-23 according to the page
#
GWG
When I was mentioning use-cases, I was talking about this section https://microformats.org/wiki/h-feed#Use_Cases
#
[tantek]
that whole section looks out of date and likely needs to be incorporated into the IndieWeb wiki instead (removed from the spec)
#
[tantek]
GWG, if you want to move it from there to /h-feed, go for it
#
GWG
Wouldn't that be the same with Examples in the Wild?
#
GWG
Only cite the wiki?
#
GWG
[tantek]: The Readers section is dated. Would that also be best moving out of the spec and just citing the /h-feed page?
#
GWG
At the moment, the /h-feed page only has publishers, not consumers
#
[tantek]
yes, worth moving Readers section to the IndieWeb wiki and linking to it instead
#
GWG
Better?
#
GWG
[tantek]: Where would you put those two lines of publishing advice about multiple feeds?
#
GWG
I'm also looking still at the first question on Zegnat's list, "Can I subscribe to tantek.com?" and thinking back to the Microsub popup earlier this year, when we discussed the idea of follow people over feeds.
#
GWG
This is looking at /h-feed at the same time as the Microformatspage.
#
GWG
Whole question should go in favor of the Discovery piece
#
[tantek]
Re: publishing multiple feeds, what does the JSONfeed spec say?
#
GWG
Only references is as "The title attribute is optional, but can be useful to differentiate multiple feeds — for instance, there might be a feed for all blog posts and a feed for comments on a specific post. These should have different titles."
angelo joined the channel
#
GWG
Looking through the JSONFeed issues list... https://github.com/manton/JSONFeed/issues/89
#
Loqi
[jbrayton] #89 Links to identical feeds in other formats
alex_, strugee and [fluffy] joined the channel
#
capjamesg[d]
[tantek] Going back to the h-feed Discovery section. I saw this: " If an implementation needs only one h-feed, it should take the first one found in any of the above methods" I would love to go a bit deeper onto this.
#
capjamesg[d]
In fact, I don't have a question. I was just reading the chat transcript and you answered my question.
#
capjamesg[d]
I haven't seen any sites that mark up two h-feeds on the same page. I can see why someone would do this (i.e. a page that shows "latest blog posts" and "most popular blog posts") but if this is not happening then there might not be a need to standardize it.
#
capjamesg[d]
GWG I commented on your implied h-feed ticket. I love the idea (and it's something I'd gladly implement in IndieWeb Search during recrawling) but I haven't seen any examples in the field of a page that has multiple h- objects but no parent h-entry.
#
capjamesg[d]
[tantek] GWG I agree that h-feeds should not just be h-entry objects. I can see myself moving my /recipes page to a list of recipes marked up with h-recipe. And using h-event markup on my /events page under a h-feed.
cek joined the channel
#
cek
k
angelo, hendursa1, rommudoh[m], reed, hala-bala[m], Abhas[m], SamWilson[m], Lohn, astralbijection[, vikanezrimaya, diegov, unrelentingtech, ChrisHarris[m], npd[m], KatMMoss[m], LaBcasse[m], nekr0z, benatkin and mackeveli_ joined the channel
#
[KevinMarks]
Per Post comment feeds are fairly common in rss/Atom, but that may be an artefact of nesting not working there. You could markup the nested h-cite comments on a post as an h-feed with an id to emulate that with minimal overhead.
[chee], tetov-irc, ShinyCyril, akevinhuang, [schmarty] and hendursa1 joined the channel
#
@QuinnyPig
Time to put on my Cloud Economics Pants and do a bit of math around @Cloudflare's R2 pricing model as described herein. https://blog.cloudflare.com/introducing-r2-object-storage/
(twitter.com/_/status/1443028078196711426)
#
[snarfed]
^ yuuup AWS's bandwidth pricing (primarily egress, but frankly all of it) has been astonishingly high for a long time. it was the main reason I migrated huffduff-video's storage away from it to BackBlaze. literally something like 1/20th the cost!
#
[snarfed]
most cloud IaaS has seen a pretty aggressive price war for a while now, that was an odd exception
alex11, hendursaga, cek1, [manton]1, [jacky], [fluffy], [calumryan], kimberlyhirsh[d], Seirdy, tetov-irc, [timothy_chambe] and marksuth joined the channel
#
@BillSeitz
↩ī¸ Promising - aligns with my use of twitter-search-links instead of WebMentions....
(twitter.com/_/status/1443304591404277762)
KartikPrabhu and plindner[m] joined the channel
#
plindner[m]
Twitter takeout gets you a blob of JS that you can massage to json... Then...... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/1ff592898a933dd2a38f6f5dcd1e37b185a868b7)
#
plindner[m]
super hacky, but I can edit the output file and get what I want. Bonus, can filter out friendfeed and gowalla stuff, lol.
barryf[d], tetov-irc, chrisaldrich, [chrisaldrich], hendursaga, [timothy_chambe] and [calumryan] joined the channel