[aaronpk], [Slackbot]1 and [schmarty] joined the channel
#[schmarty]grantcodes++ i would love to know more about this share target for together!
#Loqigrantcodes has 30 karma in this channel over the last year (45 in all channels)
#[Slackbot]1grantcodes has 30 karma in this channel over the last year (45 in all channels)
[grantcodes] joined the channel
#[grantcodes]Uhh well it's kind of a mystery how it works tbh, I think it is chrome only for now and it needs to be installed as a pwa. But I can definitely share to it on android so it works, but I doubt it currently works on anything apart from android
#[Rose]That is every IWC I've been to, I have hackday plans and then stuff happens 😛
#[cleverdevil]Hey, I am realllly happy with what I ended up working on, but it definitely wasn't anything that I had actually planned going into the summit!
valuemachine joined the channel
#[Rose]Oh same, I never regret it. (Well, Düsseldorf ended up producing a lot of work, but aside from that...)
#[eddie]I think the majority of it should be possible
#[eddie]the lack of "Collection View / Grid view" might be the only challenge. But I think I saw a work-around
#[eddie]SwiftUI is going to make things a LOT simpler. I spent so much time fighting with data and UITableView
#[cleverdevil][grantcodes] hmm, I just noticed that in "classic" layout in the latest Together, the "previous | next | close" bar at the bottom of the view obscures the actions (like, repost, etc.).
#[tantek][cleverdevil] I too ended up working on something I had not planned to do so. It really is interesting / unpredictable how the various ideas from day 1 of summit inspire us. Or even the morning / lunch conversations at day 2!
#[cleverdevil][tantek] Totally agreed, and its been true for me every single year. Two years ago, Grant and I prototyped Together on a whiteboard, and that became our project. Last year, a conversation in the hallway track with Chris Aldrich inspired me to create Indiepaper on hack day.
#[cleverdevil]This year, it was @GWG's "follow that bird" session that caused me to build out my /following page.
#[tantek]which then caused Aperture changes right?
[grantcodes] joined the channel
#[grantcodes]Hmm can you not just scroll to see the actions [cleverdevil]?
#[cleverdevil][jgmac1106] but you got it working! I saw the evidence!
#[jgmac1106]end of the day I was like.....oh yeah...I totally wanted to do a session on productivity and make a to-do list easier for me....ended up chasing an extra brace trying subgrid....
#[jgmac1106]its strange going to miss the constraint of matching character count as a creativity tool but also will have more content time for less code at same time
#[jgmac1106]gonna be in my wrap up post on Bracing for the IndieWeb...will use that as the arching metaphor
#[jgmac1106]best was trying to demo on Chrome.......totally forgot that
#[jgmac1106]mJordan is a design expert. Traded her a lesson on subgrid...well watched her learn and used her CodePen for working mf2 in her Gatsby site
#[grantcodes]Hmm [cleverdevil] yeah it does look like it is fixed instead of sticky so doesn't have the needed padding. Kinda weird it doesn't work. Looks good on macos though
#[tantek]My experience on hack day was quite unique from past hack days. While working on implementing displaying a richer reply context seemed like a "can do in a couple of hours, especially with consulting people for quick questions", I decided to double-check my approach to storing the information (from a flat file perspective) by first checking what others had done/built (and why).
[tonz] joined the channel
#[tantek]I think I knew [schmarty] also had a particular approach to storage so I asked the questions that I thought would reveal answers either reinforcing my approach, provide sensible alternatives, or new information questioning my assumptions.
#[tantek]That led to a higher level discussion about needing to partition (quarantine) information differently depending on where it came from, how (due to your own action, or from someone else's action), and the potential complexity of it
#[tantek]came up with four levels (which should likely be rethought) of sources of external information.
#[tantek]1 due to your own action, something you get a chance to see/preview first (before storing) - like previewing the information in a reply context
#Loqitantek has 20 karma in this channel over the last year (124 in all channels)
#[tantek]2 due to external action (like receiving a webmention) and a simple "response" (like, repost, reacji) where the "content" is likely innocuous (at worst a random emoji)
#[tantek]3 due to external, and richer information about the response like author name (could have offensive text, how big? needs sanitizing?) and author image (could be offensive image, big size etc.)
#[tantek]4 due to external, and arbitrary content e.g. comment/article markup
#[schmarty]my project-day ended up coming out of this discussion, and the idea that you should be able to approve whatever you're about to post, including any /link-preview or /reply-context info.
#[schmarty]b/c quill's reply feature showed a preview from XRay, but on submit only sends a in-reply-to as that URL.
#[schmarty]so i made the preview contenteditable and added handling to pack it up as an h-cite in the (JSON) micropub POST.
#GWGMy interest to start, as I already allow editing previews, is addressing this automatically
#LoqiIt looks like we don't have a page for "unlock protocol" yet. Would you like to create it?_x (Or just say "unlock protocol is ____", a sentence describing the term)
#[tantek]Out of that discussion of where to store your own content (stuff you create), vs stuff of other's you reference, e.g. according to various quarantine levels 1-4 depending on how "bad" it could be / damage it could cause, we refocused on what were the possible problems with reply-context information that might motivate changing that information
#Loqi[Tantek Çelik] #IndieWeb Summit:
reply-contexts should be user-editable in posting UIs
in-reply-to h-cite property: reason, action
* name+summary: too long, truncate
* dt-published: 1969 or future or wrong timezone, use present time
* author name: Twitter names/pro...
#[tantek]as soon as "user editing" is added to the flow of information, it becomes something you want to save/own for yourself, something of your creation that you typically can't recreate by clearing a cache and rebuilding it from the web (or even archive.org)
#[tantek]thus from a storage perspective, it starts to look much more like something you create rather than something you retrieved, which means for a reply-context, storing it along with your reply, rather than in a side-structure or side-file. for static storage that uses microformats, that would mean an h-cite for the reply-context, similar to the how to for publishing: https://indieweb.org/reply-context#Markup
#[tantek]while I was writing up those use-cases, as [schmarty] noted, he went off to actually build it and get it working in Quill!
#[tantek]looping back to the start, I was (re)considering my approach to storing the [reply-context] information. With just the primary use-case of presenting reply-context information, it made more sense to store it somewhere on the side other than my posts (per the Level 1) analysis above, though other possibilities would be ok.
[jgmac1106] joined the channel
#[jgmac1106]where do we capture everyone's project day descriptions on the wiki..or better yet where do we collect the blog posts where people reflect on their project day demos?
#[tantek]Adding the use-case of editing the information changed enough about the nature of what is being stored that it now became important to store it along with the posts! This was my original plan "save that reply-context info inside post storage alongside links to POSSE copies" but for insufficient reasons (convenience of implementation) that ignored the challenge of information from other sources, which was why I wanted to question that plan by ta
#[tantek]I bring this up in dev because ultimately I'm thinking through the challenges of "how to store stuff from other people's sites" which is a key underpinning of handling all sorts of protocols like Webmention and also all the peer to peer user interaction scenarios
#aaronpkhm did i ever write up the mess i made of storing external content in my previous website?
#Loqiaaronpk: sknebel left you a message 1 day, 10 hours ago: we have new AutoAuth issues, would love feedback on those / if you have time for a virtual session on AutoAuth e.g. on one of the next few weekends I'd be interested in running one
#[tantek]including the more challenging scenarios like dealing with abuse etc. which I think need to be thought through all the way down to the storage level
#[tantek]oh no I don't think so [aaronpk]! that wasn't the indiearchive stuff was it?
#aaronpki ended up filling like 4gb of external content, and decided that was just not sustainable
#[tantek]how we partition the storage of things makes a big difference about how it can scale, prioritizing what you back-up (since nothing is "instant") etc.
[schmarty] joined the channel
#[schmarty]heh, we have been talking recently at adafruit about ways to share data between our sites like the shop (buy stuff) and learn (docs and howtos) and i have been pushing mf2
#[schmarty]this past week the main dev for learn got to looking at it and is not excited about the use of html class attributes 😅
#[schmarty]so i told him there's an alternative, as long as we use a parser that supports it (the use of `property` that we discussed and that somebody integrated at 2018 Berlin)
[cleverdevil] joined the channel
#[schmarty]pretty sure that was aaronpk adding it to php-mf2
#[schmarty](he also lamented that it wasn't a separate chunk like a JSON island, haha)
#[tantek]it predated the current JS-JS-JS trendiness among programmers but comes from the same place. ew, mixing data in markup which is "just" a print output for display
#[schmarty]i intend to keep pushing mf2, so we'll see how i manage.
#[tantek][schmarty] ask your "main dev" what's the oldest personal text content he has stored (outside of mbox email archives)
#[tantek]systems that are storing actual content in opaque JS-accessed databases are likely to have that data disappear over time, leaving nothing but those aforementioned JSON-island summaries for people to try to discern WTH a page was about
#GWGRe the external content subject... I store them in the post as h-cites,. but I have been thinking of storing them in a linked data structure for better querying
#[tantek]GWG, not sure how "storing them in a linked data structure" makes for "better querying". I think that's one of those RDFmythisms
#[tantek]GWG, in general, consider avoiding the phrase "linked data" and the term "semantic" (especially when uppercase "Semantic") unless you're deliberately trying to refer to RDFisms
#GWGI didn't realize that is what linked data meant
#[tantek]"linked data" and those other terms are shibboleths for RDFisms
#[tantek]wow the Google dfn kinda nailed it: "a custom, principle, or belief distinguishing a particular class or group of people, especially a long-standing one regarded as outmoded or no longer important."
#[tantek]GWG, one way of partitioning information as I described (original, then external levels 1-4) is the use of separate database tables
#[tantek]I'm also considering the case of not wanting to store *any* reply-context information beyond the URL because all of the specific info is not something I want in my storage / archives. E.g. confrontational replies to bad actors
#[tantek]which is then a minor step towards dealing with much more toxic or abusive online interactions
#[tantek]looking at markup examples, where did dt-accessed come from?
#Loqi301 (301 Moved Permanently) is an HTTP status code returned from a web server to indicate the URL is permanently redirected to another URL https://indieweb.org/301
#chrisaldrichDo we have a more generic web-specific page for redirect? Surely amidst all the CMS changes folks have gone through, there would be more documentation than ^^
#chrisaldrichI just got an email from Pixabay.com asking me to fix a URL on my site to redirect, presumably because they changed their URL design (or CMS)...
#chrisaldrichThey actually wrote "Updating to the new URL avoids broken links, which can be a bad user experience and also harmful to your site's relationship with Google."
#chrisaldrichI think the bad user experience is having their staff track down all the sites using old links and asking them to update.
#Loqi[sknebel] #3 use guaranteed invalid redirect_url
#[aaronpk]yeah I'd go with either .invalid or something that is actually helpful like mine
#jgmac1106[m]But wonder if a Micropub client could be created drop in url, select dt-published range. Parse post for reply, the reply for a reply, etc publish one h-entry with child h-entries
#[tantek]Like an indieweb version of a thread unroller?
[snarfed] joined the channel
#[snarfed]ideally salmentions would make that unnecessary, but only if receivers display them all
#[tantek]I think the point is a better display, not whether they are displayed
#[aaronpk]i also forgot to describe the actually cool part of that at summit, it's more than just lazy-loaded, it's only loaded if the image comes on screen
#[snarfed]yes! definitely nice. i use the same jetpack feature on my site
#[snarfed]can't wait for browsers to support <img loading="lazy"> so we don't need JS
#[aaronpk]yeah! also cause it can be smarter about loading them just before they become visible instead of the way mine works with triggering as soon as they become visible
#[snarfed]right. that flash of broken image(s) is minor but annoying