#dev 2021-10-07
2021-10-07 UTC
# 
[tantek] I find this particularly frustrating as well. Kinda wish there was an easy path to "here's how to use your iPhone/iPod/iPad-mini to create & update your own projects for your own site" https://twitter.com/Pinboard/status/1445890223846526986
# 
@Pinboard And part of it is definitely that everyone today carries an incredible computer in their pocket and... you can't do anything with it. My first computer program at age 8 made an IBM desktop write profanity in an infinite loop. I wouldn't begin to know how to do that on an iPhone. (twitter.com/_/status/1445890223846526986)
[tw2113_Slack_] and edburns[d] joined the channel
# 
aaronpk can you make this curl request to your token endpoint and paste the result? curl -i https://example.com/token -H "Accept: application/json" -H "Authorization: Bearer XXXXX"
# 
@aaronpk ↩️ Dynamic Client Registration, but afaik no major provider supports this because they *want* RPs to have a pre-established relationship.
We built IndieAuth to avoid the need for any client registration and it works great for that use case: https://aaronparecki.com/2018/07/07/7/oauth-for-the-open-web (twitter.com/_/status/1445944383933595651)
gRegor, [tantek]1, IWSlackGateway, [snarfed]1, hendursaga, kogepan, jjuran, hendursa1, [pfefferle] and tetov-irc joined the channel
# 
capjamesg[d] I have noticed that some sites have feeds on all pages.
# capjamesg[d] I wonder how I can determine "representative feeds" that will capture most of a site's new content without reindexing individual posts on account of their being linked to a feed (i.e. comments feed).
# capjamesg[d] The homepage is the obvious way to go.
# capjamesg[d] But this would mean that likes, etc. might not be discovered as easily.
# 
nekr0z My site might be a problem. I have several homepage feeds (h-feed as a separate page because I couldn't fit it into the homepage markup, and 4 RSS feeds for different languages/combinations), every category and tag has its own RSS feed (in case someone wants to subscribe to just that; 4 feeds actually, the language variants), and also every page has a feed of comments. Should be a nightmare for your crawler :-D
# capjamesg[d] nekr0z The crawler has found at least 5,000 feeds today 😄
# capjamesg[d] That's why I raised this.
# capjamesg[d] There's also the matter of choosing which format is representative.
# capjamesg[d] I have an RSS and h-feed on my home page. They will both mark up the same data. Which should be considered representative? Otherwise I would be crawling two different feeds for the same content.
# capjamesg[d] Thus increasing crawl time and load for the site being crawled.
# capjamesg[d] As it stands, the crawler has found 1 feed for every 10 pages.
# capjamesg[d] There are contingencies for that 🙂
# capjamesg[d] (namely, crawling sitemaps and, although this has not been coded yet, scheduling full recrawls)
# 
[KevinMarks] h-feed can be richer, but if you have backcompat parsing you'll get wordpress mf1 ones which can be a bit thin. You may need to compare them on first crawl and pick a winner
# [KevinMarks] a lot of technorati secret sauce was correlating the feed version with the html version to decide which was richer. You can also have the case where the h-feed or atom/rss feed is summary not content, where you may need to crawl the post page for full.
# capjamesg[d] I forgot about the WP API. Which is worth considering due to the number of sites that use WP.
# [KevinMarks] also atom and h-feed can tell you if they're summary vs content, but rss uses description for both so you have to guess
# capjamesg[d] There is no backcompat for mf1. Should there be?
# capjamesg[d] That makes sense.
# capjamesg[d] > "correlating the feed version with the html version to decide which was richer. "
# capjamesg[d] What matters right now is just retrieving URLs. The crawler doesn't ingest data directly from XML / JSON Feed / WebSub.
# [KevinMarks] if you're using the python microformats parser it will construct a minimal h-feed from wordpress mf1 markup
# capjamesg[d] Oh that is cool.
# capjamesg[d] I am using mf2py which looks to support mf1 too.
# capjamesg[d] Amazing!
# [KevinMarks] right, but that can be confusing as the fallback depends on the theme - hentry is kinda vestiagial in WP
# capjamesg[d] That makes sense. I think some backcompat is a nice to have right now. The engine parses a lot of mf2 so adding mf1 support would take time.
hs0ucy and akevinhuang joined the channel
# 
[snarfed]1 aaronpk re Dynamic Client Registration, Mastodon does its own thing, not that, right?
# [snarfed]1 :thumbsup:
gRegor and jjuran joined the channel
# capjamesg[d] Ah no 😦
# capjamesg[d] What is your domain name?
# capjamesg[d] In terms of code, it's easy to fix that. But not without doing a full recrawl 🤦
# capjamesg[d] The code does need fixing 😄
# capjamesg[d] This is related to the discussion on representative h-feeds. I shouldn't choose a h-card if it is in a h-feed.
jjuran joined the channel
# nekr0z Actually, it's not the first comment it gets, but the first like, i.e. the first face in the facepile. It ignores the fact that every post has a proper p-author in the h-entry, and instead takes the first u-author from the facepile (and that happens to be inside h-cite u-like, so is should definitely not count as the post's author no matter what).
# capjamesg[d] I have disabled the feature where h_card_domain != domain name.
# capjamesg[d] I think I'm actually choosing the first h-card on a page.
# capjamesg[d] If none exists, I revert to the home page h-card.
# capjamesg[d] (if there is none on the home page, then the value for h_card is set to none)
# capjamesg[d] Your icons will appear, just after I recrawl again 🙂
# capjamesg[d] The issue was that I relied on the first h-card on a page other than the home page, even if, as you said, it was in a u-like-of, etc.
# capjamesg[d] Great catch! nekr0z++
# capjamesg[d] Yeah, indeed!
# capjamesg[d] I need to do a bit more debugging...
# capjamesg[d] Search is hard 😦
# capjamesg[d] But fun 🙂
# capjamesg[d] I have the spec open now.
# capjamesg[d] Thanks for reminding me about that.
# capjamesg[d] Oh, really interesting... I'll adhere to this spec in the logic.
# capjamesg[d] Yeah. The author should be linked rather than "guessing" by choosing the first h-card.
# capjamesg[d] Especially considering, say, the h-card author could post their h-card in the footer, below all the comments.
[pfefferle] and gRegorLove_ joined the channel
# [KevinMarks] Well, if its inside the post it's representative for the post, just not the site
# [snarfed]1 we have separate authorship vs representative h-card algorithms, right? https://indieweb.org/authorship
# capjamesg[d] nekr0z You can contribute a solution to the engine if you like 😄 https://github.com/capjamesg/indieweb-search/blob/main/crawler/add_to_database.py#L35
strugee_ joined the channel
# [tantek]1 snarfed++ thank you
# [tantek]1 also the embedded image here is a good case for Micropub (and "just text files on my computer") https://twitter.com/RobertHaisfield/status/1385228672580218882
# 
@RobertHaisfield @bmann @codexeditor Re why people like using multiple editors: https://pbs.twimg.com/media/EzlSE1wVgAEt-lD.jpg (twitter.com/_/status/1385228672580218882)
# nekr0z capjamesg[d]: https://indieweb.org/authorship-spec - I wasn't aware this existed, but apparently it does!
Zegnat, Guest2366 and unrelentingtech1 joined the channel
# capjamesg[d] nekr0z 🙂
# capjamesg[d] +1 [snarfed]
# capjamesg[d] nekr0z and there is a test suite!
# capjamesg[d] Amazing!
# capjamesg[d] I'll add support for this.
# capjamesg[d] Great find!
# capjamesg[d] What is an "author-page"?
# 
Loqi It looks like we don't have a page for "author-page" yet. Would you like to create it? (Or just say "author-page is ____", a sentence describing the term)
# capjamesg[d] nekr0z I saw it was italicized in the spec algorithm.
# capjamesg[d] "if there is no author-page and the h-entry's page is a permalink page, then ..."
# capjamesg[d] Thanks!
# capjamesg[d] nekr0z I just wrote some code that complies with the spec. Thanks for helping!
# capjamesg[d] I love the .rocks test suites for indieweb technologies.
# capjamesg[d] aaronpk Do you have any intentions to make a microsub.rocks service?
# capjamesg[d] Thank you nekr0z! Feel free to file as many bugs as you find! (and if you feel like the search engine is missing something that you want to see, let me know!)
# capjamesg[d] I want to make it as "indieweb" as possible 🙂
# capjamesg[d] What is teuxdeux?
# Loqi It looks like we don't have a page for "teuxdeux" yet. Would you like to create it? (Or just say "teuxdeux is ____", a sentence describing the term)
# hs0ucy At least the year is not 1970 ^^
# capjamesg[d] My blog home page?
# capjamesg[d] That's odd haha.
# capjamesg[d] Thanks for sharing aaronpk.
# capjamesg[d] Ah, I was using a capital M for month when it should have been lower acse.
# capjamesg[d] *case
# capjamesg[d] angelo++
# [tantek]1 that sounds like datetime sprintf codes or something
# hs0ucy https://imgur.com/wRYv0Mh.png
# capjamesg[d] I just fixed that issue.
# capjamesg[d] Thanks everyone!
marksuth1, [schmarty]1, [tw2113_Slack_]1, IWSlackGateway3, [snarfed] and [fluffy] joined the channel
# capjamesg[d] nekr0z authorship code is now public on GitHub. Thanks again!
# capjamesg[d] Yep 🙂
# capjamesg[d] I use Typora for almost all of my non-dev markdown (i.e. adding posts manually to my site in my IDE). They render quite a bit of HTML but tbh I don't like that feature.
# capjamesg[d] That makes sense.
# capjamesg[d] In my Jekyll case, I am more using .md in the spirit of front matter. I actually wonder if I could use a .html file instead.
# capjamesg[d] I think Jekyll might support .html and front matter as part of their parsing.
# capjamesg[d] (But all of my blog posts are in almost completely pure markdown because, well, I wrote them in markdown)
# capjamesg[d] ins del?
# capjamesg[d] Wow! I had no idea they existed sknebel.
# capjamesg[d] I done a quiz on semantic HTML a few weeks ago. It was fun.
# capjamesg[d] I like Jekyll's suggested layout because it is clear to me. Front matter is visually and programmatically distinct from the page content.
# hs0ucy That's why i'm not a fan of front matter ... it's easy to extend a SSG to other markup languages without it
# capjamesg[d] I can see the convenience factor in that.
# capjamesg[d] Especially for common things.
# capjamesg[d] Am I right in saying you can actually make custom HTML elements?
# capjamesg[d] I think I saw this somewhere.
# hs0ucy yep
# capjamesg[d] I like my markdown as simple as I can get it 🙂
# capjamesg[d] [fluffy] I don't have any use case for it.
# capjamesg[d] But it's interesting.
# hs0ucy capjamesg[d]: https://developer.mozilla.org/en-US/docs/Web/Web_Components/Using_custom_elements
# capjamesg[d] That's what I was reading.
# capjamesg[d] I have mixed feelings on that. On the one hand, there is flexibility. But on the other, it makes it harder to distinguish semantic HTML from custom HTML to those who haven't done much with HTML.
# hs0ucy yep
# capjamesg[d] I haven't used it so I can't say it is bad or good. Those are my thoughts right now, however.
# capjamesg[d] +1
# hs0ucy it's more for app or templating ... component
# capjamesg[d] Semantic HTML is rich in itself. I can't think of a moment where I thought to myself "what I need here is a new HTML element..."
# capjamesg[d] That makes sense hs0ucy.
# hs0ucy capjamesg[d]: HTML have ~130 elements already ^^
# capjamesg[d] Yep!
# capjamesg[d] More than enough!
# capjamesg[d] (for most use cases, I'm sure there are edge cases)
# hs0ucy i agree
# hs0ucy Anyway with tables and divs you can do everything :P
akevinhuang2 joined the channel
# sknebel ah right, https://soupault.app/
# hs0ucy sknebel: it's an SSG or what?
# [KevinMarks] The new html element I want is <playlist >
# capjamesg[d] What would that do [KevinMarks]?
# [KevinMarks] And yes I could in theory use SMIL but no.
# capjamesg[d] > Anyway with tables and divs you can do everything 😛
# [KevinMarks] It would play a series of <audio> or <video> elements without gaps
# [KevinMarks] Instead we have HLS. Sigh.
# hs0ucy [KevinMarks]: SMIL is deprecated no?
# [KevinMarks] Right. I mean I could use it to make a playlist but then I'm writing 80s playlist formats in a separate file.
# hs0ucy [KevinMarks]: multiple <audio> elements in <ul> ? :P
hendursaga joined the channel
# [KevinMarks] Without gaps is the hard part. You really need to be the browser to do that right. You can do it with low level audio api but that's a pain.
# hs0ucy You mean sound gaps?
# hs0ucy sknebel: there's good ideas in soupault
# hs0ucy maybe I'll borrow some for my ssg
# [KevinMarks] Sound gaps, but Video gaps to. To avoid Sound gaps you need to be sample accurate, which needs native code
[Ed_Beck], voxpelli_, Allie, chenghiz_, willnorris, tetov-irc and gRegor joined the channel
# 
@samuelgoto ↩️ @aaronpk on a related note: does any part of IndieAuth break when browsers block third party cookies? (twitter.com/_/status/1446259167178608640)
hs0ucy joined the channel