#[manton]Morning, IndieWeb folks! 🙂 I’m finally wrapping up my book Indie Microblogging. If anyone has any feedback, happy to hear it… Most of the IndieWeb details are in Part 3 and some in Part 5. You can skip the contents here to get an idea: https://book.micro.blog
jacky_, [jeremycherfas], retronav_, 011AAHWMZ, angelo, epoch, kushal, wagle, lagash_, superkuh, jeremycherfas, tetov-irc, s[_], srushe, oodani, vikanezrimaya, chenghiz_, Saphire, sebsel, rrix, Xe, jbove, Feoh, klez_, Demi, voxpelli, BinarySavior, IWDiscordGateway, [Murray], jacky, barnaby, max and [KevinMarks] joined the channel
#capjamesg[James_Van_Dyne] How are things going with the indieweb-utils doc restructure?
jjuran, gerben and [timothy_chambe] joined the channel
#vikanezrimayaIs it generally reasonable to set a long Cache-Control: max-age on media endpoint uploads even if the underlying storage/URL scheme is not content-addressable?
#vikanezrimayaI thought of setting very aggressive caching headers on my media endpoint, but since the backends are actually somewhat pluggable, I can't reasonably assume that the underlying storage is content-addressed.
#vikanezrimayaThe only requirement AFAIK is to have URLs be unguessable to an external observer, which in combination with the intent of the media endpoint means URLs are sufficiently unique and long-lived to justify hyper-aggressive caching.
#[schmarty]the major question there is whether the contents at a particular URL will change. content-addressability is only one way of ensuring the contents at a URL will change.
#vikanezrimayaAre there media endpoints that allow "replacing" a file? Mine, for example, doesn't, even with the default CAS backend.
#vikanezrimayaI don't think there is a standard method in the Micropub protocol for "replacing" a file, since I don't remember media endpoints supporting, say, PUT requests (which AFAIK are an idiomatic way to unconditionally replace something if it exists)
#[schmarty]i don't believe the standard specifies any rules for where a Micropub media endpoint "puts" a given file in terms of the URL.
#vikanezrimaya[schmarty]: it does -- the URL must be unguessable. That's the only rule.
#[schmarty]so nothing about avoiding collisions or deduping or etc.
#vikanezrimayaFor me it's unguessable to an external observer who can't sniff traffic, since the underlying storage is content-addressed. Dupes are handled... I believe by actually not handling them at all, CAS solves that problem entirely
#[schmarty]same with /Morris ! choosing to make content-addressable URLs saved a lot of thinking.
#[schmarty]but the question of whether to send long-lived cache headers is more about whether a given micropub media endpoint storage backend promises that a given URL's contents are stable. ¯\_(ツ)_/¯
#[schmarty]_how_ the backend guarantees that doesn't really matter.
#vikanezrimayaI guess I'll use etags instead! They're supported by a lot of places!
#[tantek]4been deep-diving a bit more on considerations for post authoring when you plan to POSSE to Twitter (and thus only the first ~256 chars of your post will be POSSEd)
#[tantek]4It's a mix of content-authoring considerations, and how a good authoring UI could help with those considerations
#[tantek]4for example, you want that first 256chars to be "interesting" enough that it encourages someone to click-thru to read your full post, and also clearly communicate your overall topic well enough to minimize people clicking through and being disappointed or surprised they didn't get the content they wanted/expected (that kind of click-thru is a very short term thing, you don't want people to associate clicking thru with
#[tantek]4another example: since a key point of POSSEing is distribution and /reach, at least in a positive way to friends who may still be "stuck" on the silo destination, front-loading things like hashtags and @-mentions in that first 256chars makes a lot of sense
#vikanezrimayare: Twitter POSSE (side note): back when I still posse-d I noticed twitter heavily penalizes posts including links (which I have been doing to have backlinks to originals), which also impedes discovery, e.g. via feeds
#[tantek]4an @-mention in the first 256chars will (likely) notify that person of the @-mention, whereas an @-mention later on in your post won't result in any such notification (except in the rare cases where your post links the @-mention to their personal domain instead which accepts webmentions which your posting code sends)
#vikanezrimayaalgorithmic feeds are a menace upon society and must be extinguished once and for all
#[tantek]4Twitter also has an option to turn-off their algo feed and show purely reverse chronological instead
#vikanezrimaya[tantek]4: yes but that's per-user and also reverts back after a while
#vikanezrimayayou can't really opt out of looking at the algorithmic feed, it'll return. And you can't opt out of being de-ranked by the algorithmic feed, since that kinda defeats the purpose of the algorithm
#[tantek]4has not reverted back for me in many many months (over a year) in the iOS native client
#vikanezrimayaI remember the feed xplicitly telling it'll eventually revert back... maybe I haven't visited twitter in a long time
#vikanezrimayaI actually prefer RSS nowadays, tho I do want to eventually impl microsub and migrate to that
#[tantek]4it did used to do that (revert) a long long time ago
#[tantek]4maybe this would be a good time to start a campaign to get Twitter users we know to change their home timeline preference to "Latest Tweets"
#[tantek]4I have also heard that tweets with links get deprioritized in the algo feed. the POSSE brainstorm workaround for that has been to *not* include the original post permalink in the POSSE tweet, and instead auto-reply to the POSSE tweet with a reply-tweet like "@-self originally posted at (post-permalink-URL)"
#[tantek]4that way you get more algo feed distribution for the original POSSE tweet, and you even get the full 280chars as well
#[tantek]4the one downside to this is for anyone following you, you "take-up" more space in their home timeline view because Twitter often does show one-reply (especially if you are following both people) to a tweet, even in composite stream view
#[tantek]4I've thought about experimenting with it anyway just to see if people (on Twitter) find it annoying or not
#gRegorcapjamesg++ good post, interesting as someone who hasn't participated in CGs
#Loqicapjamesg has 31 karma in this channel over the last year (92 in all channels)
#sknebelI keep wondering what that name reminds me off, but of course, of dustyweb, cwebbers (old?) handle :D is she involved?
#[tantek]4domain name << beware of domain names similar to your own being squatted, or of unintentionally linking to squatted domain-name typos instead of major media sites (e.g. links from social media) https://www.disinfo.eu/doppelganger
#capjamesgAt the beginning I was kind of worried about not having much direct experience in the subject matter but I left feeling like I could contribute. That was empowering.
#LoqiThat! (or "that ^" or "that ^^^") is a rarely seen reply often emphasizing agreement with a This post, but sometimes[1] merely emphasizing agreement with a previous reply https://indieweb.org/that
#[James_Van_Dyne][jamesg483] I haven’t had the chance to look at it quite yet – my son’s been waking up early these past few days, which eliminates my focus on the computer time. Looks like the current docs aren’t building for me locally, so debugging that now.
#capjamesg[James_Van_Dyne] I think I know the solution.