trwnhyeah gRegor i would likewise assume that only the top-level .u-in-reply-to would affect the display of the post... i'm just trying to get the semantics right as a producer, and i'll let consumers figure out whatever they wanna do with it
trwnhhonestly the main thing i seem to be drawing from this is that the concept of "post types" or "post kinds" doesn't actually make a ton of sense, especially for producers
[tantek]🌉 Admin notice: Due to a Bridgy Fed bug, @tantek.com's "ActivityPub actor" was deleted, notifying every following server which then unfollowed @tantek.com. It was an accident! Feel free to refollow @tantek.com at your discretion. Sorry for the inconvenience!
[tantek]I think that covers the key details, and what users might expect to get / be ok with seeing in their DMs (based on similar messages I've seen on email etc.)
[tantek][snarfed] if you do plan to send out such (semi-?)automated DMs, please LMK when they've been sent, and I'll post a public follow-up accordingly
[0x3b0b]For purposes of the sorting thing...I'd probably use whatever tool I had handy to make a csv, and then do the manipulation with powershell. But I'm weird.
trwnhok oddball question: does it make sense to nest an h-entry inside u-in-reply-to? i'm mostly wondering this because i want to markup a minimal representation of the thing i'm replying to, and i was thinking i could do that with p-name and u-url and nothing else.
[tantek]Not really because h-entry has an implication of a complete thing, an entry, as opposed to a reference to a thing, for which we prefer to use h-cite
[0x3b0b]I'm going to have to investigate this Tantek business this weekend. I used to follow Tantek from my website, I'm positive...now I can look up his posts and his profile, his profile says I'm at "follow request sent..." but if I try to undo the follow request so I can try again, it throws an internal server error that ends with telling me "actor was deleted." I suspect I have wound up with a weird cached state.
[snarfed]0x3b0b by "follow Tantek from my website," do you mean via ActivityPub? does your website use Bridgy Fed to federate? or something else? or am I misunderstanding?
[0x3b0b]I suspect my end has an edge case where the fact that it thinks his profile was deleted is more important to it than the fact that it is clearly _there_
[snarfed]microblogpub may not technically be doing anything wrong. I suspect (worry) that re-creating an actor after it's been deleted may have mixed results across the fediverse. it worked with lots of implementations, but clearly not all
trwnhhm. so would that make it an h-cite embedded in u-in-reply-to? or should i just drop the idea and not mark up the p-name of the thing i'm replying to at all? > [tantek] Not really because h-entry has an implication of a complete thing, an entry, as opposed to a reference to a thing, for which we prefer to use h-cite
trwnhthe reason i am thinking about this at all is because i am currently representing my replies/responses by both name and url. and then i use this information to render something like <span>in response to <a href={{.url}}>{{or .name "[link]"}}</a></span>
[0x3b0b]For what it's worth, my site has u-in-reply-to and h-cite on the same element. Here's a permalink to one of my notes that's in the middle of a conversation thread, if you want to check the source. I might not have it "right," but I have it "a way." I'm finally heading to bed though, so I am only open to discussing it...some other time.
trwnhstarting to have second thoughts so i should probably step back and try to describe what i'm trying to do here semantically. if i quote something, is that always appropriate for an h-cite? because what i'm really trying to do here is mark up my blockquotes and related metadata within a single container. this is why i'm unsure and generally asking somewhat leading questions regarding the use of h-cite, h-card, u-in-reply-to, u-quotation-of. i'm
trwnhthat is to say: if the purpose of the quote is to respond to it, then am i responding to the quote, or to the full resource? because on a top-level, my h-entry is in response to some url representing the full resource... but does that make each quote a "citation"?
[KevinMarks]The delete then recreate case is worth documenting as a possible AP spec issue, as if implementations treat delete as irrevocable that sounds like an attack vector. I can't work out all the implications, but it may be a way to disrupt account migration too
[snarfed]trwnh good questions! microformats and the indieweb have recommendations, but not necessarily always definitive, normative answers. you can put together mf2 and other building blocks in a number of ways
trwnhyeah [snarfed] the first thing i did was read those pages ^^; the main thing i'm getting out of my reading though is that a lot of the time the markup and building blocks seem unduly weighted toward recreating silo functionality and maintaining compatibility through POSSE. meanwhile i'm trying to start with a mostly clean slate approach and i'm more interested in the semantics being fully expressive. so i'm considering using both u-in-reply-to
trwnhand u-quotation-of even though [tantek] above implied that they were indeed disjoint. i'm also considering *dropping* h-cite because i'm unsure that quoting someone is always citing them. but i might use some kind of h-x-* to avoid having the properties of the quote leak into the scope of the h-entry
trwnhi guess the thing i need to think more about is, what are the semantics of .u-quotation-of? because .u-in-reply-to seems pretty clear, you are responding to the thing
LoqiA citation is a reference to something (a way to look up what) someone said, wrote, or photographed/captured, typically a published work of some kind, these days often on the web https://indieweb.org/cite
trwnhi guess what's tripping me up (aside from lack of examples as you mention) is that i think the main entity is not the link or author name, but rather the quoted content (or possibly an e-summary as we discussed earlier)
trwnhthe main use case for now is like how you'd quote someone on a forum to reply to part of their post, which i *guess* would currently best fit into .u-in-reply-to.h-cite? it's just a bit of mental weirdness because it doesn't feel like a citation to me. maybe i'm wrong about this but i see h-cite and citations as kind of a way to collect references such that you could build a list of child h-cites within the larger h-entry. at least that is what
trwnhmost of the current documentation leads me to believe, with the language calling out "citations and references" and the examples not including excerpted content (only a wee mention of p-content for tweets) and the focus being "a smaller simpler set of only 8 properties to solve the specific problem of how to markup citations in an article that refers to other articles"
trwnhlike, if the usage/purpose/focus of h-cite has evolved away from the strict subset described on indieweb and microformats wikis, then maybe that should be called out or clarified somewhere? because i am certainly going to use h-cite in marking up my essay references, i'm just unsure if it *also* applies in a non-essay non-article context where you are not really intending the quote to be added to a list of references, you are doing something
trwnhelse with it (like responding to it with u-in-reply-to, or making it into a "quotation" with u-quotation-of (which i'm still not clear what that entails semantically))
[tantek]Yeah don't overthink it. Like I said, even if you only mention a name, @-name, or URL along with a quote or larger reply context for example, it's still a citation
trwnhi wasn't around for those but i'm vaguely aware that there is some disagreement between various authorities on whether it should always refer to a creative work or if it's fine to refer to people. fwiw i generally find the former argument more convincing (per MDN, WHATWG, W3C) but i get why someone would want/think of using it on people as a means of attribution.
trwnhin any case yeah there should be more clarifying documentation that "even though it's called h-cite and started out as a way to markup citations as reference, it has evolved to be used in more informal contexts" or something like that
trwnhbut re: "identifying the source of a quotation" this is what's tripping me up still, because some quotes don't have a source, or the source is offline and therefore inappropriate for the cite= attr
trwnhin the "quote to reply in a forum" use case this is generally not the case because the quote is simultaneously a quote (with e-summary?), reply (u-in-reply-to), has attribution (the username/etc), and has citation (url of the forum post)
trwnhfor which i would remove the .u-in-reply-to since it's not a reply, but also because the quote doesn't have a citation then there is no url to refer to, and u-quotation-of feels inappropriate (because doesn't u-* mean url parsing?), and again there's no citation there's just a quote with attribution. so this leaves me with zero markup of the quote, and no h-* entity to markup the attribution either (for example with p-name)
trwnh(this is what leads me to considering `.h-x-quote .p-author.h-card` as the attribution and `.h-x-quote .e-content` (or e-summary?) as the quoted content (or summary?)
[tantek]"attribution but not citation" doesn't make sense, the attribution (in any form) is presumably a source of some kind, even if it is only the name of the person who said it
[tantek]start with using <blockquote> </blockquote> for the text being quoted, and <cite> </cite> for the attribution, that's good enough for now. inventing a bunch of one-off microformats is generally a waste of time unless you're planning to also write a bunch of consuming code with a specific consuming code use-case in mind
[0x3b0b]I feel like I get where you're coming from, because if you aggregate some of the information from various places about "citations" in general (not the Chevy), the cite element, and h-cite, I can see how they come across as things that may have drifted apart more in usage than one might expect given their relationships. It's a kind of conversation I'd likely be eager to engage with in various ways if I had more energy today, and one I might be...
[0x3b0b]And on further reflection I don't think I have any productive advice so I'm going to go back to staring into space and listening to music for a while
[tantek]like any kind of semantics, if you don't have a consuming use-case in mind, especially one that others have expressed interest in it, you're likely overthinking it, and definitely overcoding it
[tantek]stick with the best/closest plain HTML you can until you can actually describe a consuming use-case you plan to code or someone else is interested in coding
[0x3b0b] I'm going to have to investigate this Tantek business this weekend. I used to follow Tantek from my website, I'm positive...now I can look up his posts and his profile, his profile says I'm at "follow request sent..." but if I try to undo the follow request so I can try again, it throws an internal server error that ends with telling me "actor was deleted." I suspect I have wound up with a weird cached state.
trwnh hm. so would that make it an h-cite embedded in u-in-reply-to? or should i just drop the idea and not mark up the p-name of the thing i'm replying to at all? > [tantek] Not really because h-entry has an implication of a complete thing, an entry, as opposed to a reference to a thing, for which we prefer to use h-cite
capjamesg[d] This is a neat CSS rule: https://developer.mozilla.org/en-US/docs/Web/CSS/mix-blend-mode
[KevinMarks] The delete then recreate case is worth documenting as a possible AP spec issue, as if implementations treat delete as irrevocable that sounds like an attack vector. I can't work out all the implications, but it may be a way to disrupt account migration too
[snarfed] here, some of those recommendations are I don't think for reply contexts, https://indieweb.org/quotation#How_to_markup for quote posts
Loqi A citation is a reference to something (a way to look up what) someone said, wrote, or photographed/captured, typically a published work of some kind, these days often on the web https://indieweb.org/cite
[tantek] start with using <blockquote> </blockquote> for the text being quoted, and <cite> </cite> for the attribution, that's good enough for now. inventing a bunch of one-off microformats is generally a waste of time unless you're planning to also write a bunch of consuming code with a specific consuming code use-case in mind