[snarfed]aaronpk we haven't sketched out a "pure web" or "indieweb" version of backfeed, have we? ie something like an extension to webmention where servers would send to rel-canonical or similar links in targets, if found, and then receivers would also support incoming wms where the target is u-syndication link instead of the target?
aaronpkaw man i really want to write a webmention game where you have to collect monsters by sending webmentions to them, but every time i sit down to start sketching it out it starts to look overwhelming
[prtksxna]aaronpk: I was wondering, do you imagine Webmentions (and the rest of the stack) being used for games and other kinds of apps regularly? (sorry if this isn’t the correct channel, happy to move)
aaronpkso the one question here is whether the comment on dev.to should look like a comment on the original post or have any reference to the dev.to syndication
Zegnat“multiple authors is a thing” - but not very often? I seem to recall some people looking for blog examples to document and having a hard time finidng a lot? Maybe was sknebel…
[LewisCowles]I cannot imagine what it would be for. I understand multiple people being involved in a work, but my understanding of flow is that each section / article would have a single author, perhaps an editor and maybe an editor in chief
[LewisCowles]I'm nerdy interested in the merging of h-cards. I'd assume newest is last in merge to blat the others, and each can overlay something PATCH, but can equally see a case for treating it as a PUT or whole-document operation
ZegnatHad to run out for a meeting. I meant that every property has its values treated as a set. Just keep adding to it. Maybe keep track of when you added certain values.
ZegnatI do now wonder if I should have an updated property on my h-card. It gets updated every time I weigh myself, just to name one reason for its many updates
[KevinMarks]Multiple urls is very common in h-card, especially if the email and phone listed are urls too. Multiple authors are common, but less so on personal websites, because they're personal.
[LewisCowles]I suppose I'm familiar with the process for a research paper, where it's the sum of efforts, with a role for someone to add one-voice. Thanks for enhancing my awareness.
bekofup on the Cloudflare issue: Something seems to be really broken with the particular setup. Owner has no idea where this is coming from but telling me that half of the WordPress admin seems to be broken. Moving away from Cloudflare is considered already. Well if anything blocks wp-admin all the time even for the owner I can imagine there is more amiss with this installation.
bekoswentel: o/ Is caption on image upload working for indigenous-android? Did nothing for me and I wonder if this is a bug in the Micropub plugin or the app.
[prtksxna]beko: The wp-admin stuff doesn't load for me at times too, even with the page rules. Hard refreshes usually helps there, but kind of frustrating.
[LewisCowles]beko: this is definitely CloudFlare interaction with GutenBerg. I'm fairly certain they need to whitelist the json api and /wp-admin/ to remit
beko[LewisCowles]: wew. I've a hard time to wrap my head around this. Something that big messing up with a software used by such a huge number of websites nowadays.
jamietannabeko FYI you may get alt text on `mp-photo-alt` or `mp-photo-alt[]` depending on whether you send one/multiple images, so worth checking if both are handled
ZegnatI am guessing sknebel would have wanted the alt to be JSON POST only, not also in form-encoded, because it just makes form-encoded ever more complex?
bekoI never bothered much with captions. Recently we had an outage of our CDN so we got served like zero pictures at all. That was an interesting experiment so accessibility kinda popped up on my radar :D
jamietannaRe form encoded params with square brackets, I want to say over the weekend I saw that it's part of the URI spec, but can't find where I saw that - will dig for it after lunch
sknebelyep. the spec already provides a way to set alt text and from what I gather intentionally kept formencoded as the simple but limited path (which makes supporting the bits in the spec relatively easy)
sknebelit's yet another thing. I personally had enough issues of implementing the full JSON variant of the spec + basic conversion of the formencoded standard format just to find that clients don't want to speak that that I've mostly given up on MP, because adding yet another way of doing the same thing is not particularly interesting for me
sknebelconverting the normal properties from formencoded to json is luckily relatively easy (take all the params, put them in a json object, pass that to your json implementation that actually does data), so I didn't mind adding that to my endpoint that I build "json-first"
[grantcodes]Yep, I am in the pro json camp too. Also for updates and more complicated functionality it is required. Luckily in node the conversion from form encoded it pretty easy - apart from non standard properties like `mp-photo-alt`
[grantcodes]Dunno, I feel like most people who are knowledgeable enough to send a form encoded request to a custom endpoint with tokens can probably send json without too much extra learning.
sknebelright. from what I understand, the idea behind formencoded was that you can get a quick update posted etc with really simple code (potentially a standard HTML form, or anything that can send POST). and then JSON for more complex things, where you're building UI anyways
ZegnatBut HTTP has no real concept of arrays or nesting when it comes to POST payloads. How languages arrive at { a: [1, 2] } is all over the place, I think I have seen ?a=1&a=2 ?a[]=1&a[]=2 and even ?a[0]=1&a[1]=2.
[LewisCowles]when leaning on what HTTP / HTML has by default, some behaviours are not provided just by using the standards though. Check out how Ruby deals with objects that make more sense in JSON (Zegnat making the point as I type it)
ZegnatPHP does the numbered indices. But I think Flask can’t read those by default because it wants to do getList('a[]') ? Or something? I would have to re-research it, and not too keen on that :P
[LewisCowles]someone started to make a parser for flask to do it, but their response was "Why would you want that" and "for interop" was not a good enough reason
ZegnatI also like how neither RFC referenced by PHP (RFCs 1738 and 3986) include the string "urlencoded". It really is one of those situations where you just need to wish people luck :P
[LewisCowles]PHP RFC's are hilarious, but I love them. Well over 20 years at the mast and they still haven't delved into obscure choices like. What if we made a whole other way of saying if(!something)
[kimberlyhirsh], [Rose], nickodd, [bonkerfield], [fluffy], sscarfe, KartikPrabhu, [grantcodes], [snarfed], [jgmac1106], gRegorLove, cal and swentel joined the channel; nickodd left the channel
calHas anyone worked with Pipedream (https://pipedream.com/) esp. with regards to connecting to IndieAuth apps? Finally blocking out some time to implement Micropub on ThreadReaderApp, and I am considering using Pipedream as a fun quick way to do it.
aaronpksebsel: that advice was from the old implied name mf2 parsing rules that meant every object would be forced to have a "name" property. leaving off the p-name meant more often than not, the "name" property would be full of junk from surrounding elements, hence the guidance to use p-name with e-content
aaronpkbut now that the parsers have updated, (and it's been a long time), there's no reason for adding that class anymore, and now just makes more work for people on the consuming end to de-duplicate the "name" and "content" properties
swentelGetting some feedback from my girlfriend who's testing indigenous, and the only thing that she really misses is 'bookmarking'. Been breaking my head around it the last few hours, but can't really come up with something with the existing spec at the moment, so ideas welcome :)
LoqiRead later is an indication on (or relating to) a website that one wants to save the URL to come back and read the content at a future time https://indieweb.org/read_later
swentelIndigenous has an option to configure a 'read later' channel, so an item moved in there gets marked as unread automatically. But the bookmarks channel won't trigger that
swenteland it would only be an option in the drupal server then, no new spec or so for microsub, or indigenous being the only client having that feature ;)