#dev 2021-06-08

2021-06-08 UTC
[jacky] joined the channel
#
aaronpk sort of related to the IndieAuth discussion, OpenID Connect “SIOP” (Self-Issued OpenID Provider) also doesn’t have a token endpoint and instead uses the implicit flow to return user identity information
#
aaronpk i’m not really a fan of that approach but they are pretty firm on the OpenID Provider in that case not having a token endpoint
#
aaronpk one reason is they want an app on your phone to be able to be an OpenID Provider, so it wouldn’t be accessible to clients necessarily
#
aaronpk and if you haven’t connected the dots yet, this app would of course be a blockchain wallet
#
@FaebornNetworks ↩️ It works really well, except for the times when IndieAuth just somehow disappears and the account is randomly disabled. Have seen that a few times. (twitter.com/_/status/1402093436526903296)
samwilson, hendursaga, [chrisaldrich], capjamesg and [jacky] joined the channel
#
[jacky] interesting form (evolution?) of sqlite: https://github.com/rqlite/rqlite
#
Loqi [rqlite] rqlite: The lightweight, distributed relational database built on SQLite
#
Saphire Hm
#
Saphire Trying to look into making an image host but uh, slightly confused on how to deal with storage part @-@
#
Saphire aaronpk: uhhh, wait when did blockchain get involved with OID?
#
Saphire prodprods DID
#
Saphire Hm
#
Saphire ... everything related to it is uh, very hard to read as it just abstracts the abstract
hendursa1, justBull, [Murray], [KevinMarks], IWSlackGateway, hendursaga, alex_, reed, calebjasik, Abhas[m], batkin[m], nekr0z, [tw2113_Slack_], [chrisaldrich], capjamesg, gRegor, chenghiz_, jeremycherfas, aaronpk, alex11 and [jacky] joined the channel
#
[jacky] https://mastodon.social/@mbajur/106370312765086032 a nice take on a dashboard for one's site
#
Loqi [mbajur] Just a little sneak peek of a new #activitypub based project I'm working on https://files.mastodon.social/media_attachments/files/106/370/304/303/780/269/original/80edf9905172f12f.png
#
Zegnat That is a pretty design
#
@clawfire Working on webmentions on my blog. Finally I’m starting to put some style on it. Playing a bit with css grid. Nice but style need to master few stuff … will probably go back to flexbox :p (twitter.com/_/status/1402318532730363908)
[tantek] joined the channel
#
[tantek] Does anyone have (or know of) a convention for publishing "estimated reading time (in minutes)" in an Atom feed? Atom has no built-in capability, and I don't know of any Atom extensions to publish/consume "reading time"
#
Loqi [tantek]: capjamesg left you a message 4 days ago: progressive enhancement is interesting. The feature doesn't quite fall under that category because, ultimately, the button appears irrespective of whether you have JS enabled.
#
[tantek] ^ precisely
jeremycherfas and capjamesg joined the channel
#
[tantek] Hah this is great, also *classic* flaw in command line UI design (either common or difficult to avoid, thus nearly innate) https://www.theverge.com/2017/3/2/14792442/amazon-s3-outage-cause-typo-internet-server
#
[tantek] > “Unfortunately, one of the inputs to the command was entered incorrectly and a larger set of servers was removed than intended,” Amazon said. “The servers that were inadvertently removed supported two other S3 subsystems.”
#
capjamesg tantek I had a good laugh reading that a year or two ago.
#
capjamesg Super interesting that one command could do that on the scale of AWS.
#
capjamesg tantek I think this is related: instagram.com/
#
capjamesg Whoops. Not that.
#
capjamesg This: https://aws.amazon.com/message/41926/
#
[tantek] Nice find
#
capjamesg What a vague way of saying they fixed it: "While removal of capacity is a key operational practice, in this instance, the tool used allowed too much capacity to be removed too quickly. We have modified this tool to remove capacity more slowly and added safeguards to prevent capacity from being removed when it will take any subsystem below its minimum required capacity level."
#
capjamesg More the second sentence.
#
[tantek] S3 << 2017-02-28 outage summary from Amazon: [https://aws.amazon.com/message/41926/ Summary of the Amazon S3 Service Disruption in the Northern Virginia (US-EAST-1) Region]
#
Loqi ok, I added "2017-02-28 outage summary from Amazon: [https://aws.amazon.com/message/41926/ Summary of the Amazon S3 Service Disruption in the Northern Virginia (US-EAST-1) Region]" to the "See Also" section of /Amazon_S3 https://indieweb.org/wiki/index.php?diff=76022&oldid=76021
#
[tantek] imagines next step being two admins with keys they need to turn simultaneously
#
capjamesg I'd be somewhat scared being a sysadmin if I thought I had such power.
#
capjamesg Maybe there would be the scenes like in the movies with missiles. The two sysadmins would look at each other, nod, and then twist their keys.
#
[tantek] yes it was a Wargames reference 🙂
#
[tantek] "Turn your key, admin!"
#
capjamesg Oh, I've not seen that in a long time!
#
capjamesg Not directly related, from 2020: https://blog.cloudflare.com/cloudflare-outage-on-july-17-2020/
#
capjamesg I think I remember reading about that when it happened.
#
capjamesg Because headlines about parts of the internet going down always catch my eye :)
#
[tantek] is still waiting for someone to name their personal IndieWeb CMS the "WOPR"
#
capjamesg Pronounced "whopper" :D
[KevinMarks] joined the channel
#
[KevinMarks] Known shows estimated reading times, but not sure how it's calculated.
#
[KevinMarks] eg http://known.kevinmarks.com/2018/douglas-adams-on-digital-id
#
[KevinMarks] says '2 min read'
#
[KevinMarks] https://github.com/idno/known/blob/5184f510b2074d09890dfcf4f9d87875d3112fe5/Idno/Common/Entity.php#L1258
#
[KevinMarks] so estimates 200 words per minute reading
#
Zegnat Is there any reason to publish reading time in a feed? Would that not be better left to the reader? If they need a UI to display the data taken out of the ATOM anyway, maybe it is better to have them fill the UI element
#
[tantek] Zegnat, it would literally be for a /Reader, e.g. so it could display it in a visual summary (maybe an icon!) in a list view
#
[tantek] good call asking for the use-case tho 🙂
#
Zegnat Yeah, I am just wondering if you would not want the reader to calculate it. Instead of figuring out ways to send it to the reader from your end.
#
[tantek] it's not either or
#
[KevinMarks] the reader app could possibly calculate reading time empirically
#
Zegnat now wonders what it would take to train a reading time estimator
#
[tantek] point being, if the publisher/author is already providing a particular estimated reading time, display it. if the reader wants to do its own (less precise) heuristics, it can do so as a fallback
#
[tantek] per KevinMarks it could also figure out a per user&author pairing reading time "factor", so if it consistently took you say half the time that someone's reading time said, it could display that instead 🙂
#
[tantek] the use-cases are there
#
[tantek] I wasn't trying to debate that
#
[tantek] I was merely looking for existing examples of "reading time" publishing in feed files as an explicitly structured thing, not just as inline in the contentx
#
[KevinMarks] known is almost marked up as structure
#
[KevinMarks] `<span class="vague">2 min read </span>`
#
[KevinMarks] `<p class="reading">`
#
[KevinMarks] `</p>`
#
gRegor heh, class=vague
#
[jacky] trying to do some brainstorming. I got Koype working in a way where it's solely a Micropub endpoint - it stores stuff to disk with a SQLite database for some indexing of content and the likes.
#
[jacky] Right now, I'm trying to wonder what would be a 'nice' way to handle rendering of this content without changing too much of how Koype works (more of the info - in Rust - is stored here https://git.jacky.wtf/indieweb/koype-next/src/branch/develop/src/storage/mod.rs)
#
[jacky] I don't want to fuse Micropub too tightly to my site (I'd actually rather use `q=source` to fetch the content _from_ Koype to handle posts) but I don't immediately know now I'm just being too granular in the separation of concerns)
#
[jacky] My hope in keeping them separate is to make it easy for me to iterate on Micropub stuff while keeping my site working normally
#
[jacky] Right now, I'm thinking about using Eleventy to render/build my site every few minutes or so using the pagination aspect of `q=source` (and ideally by having a feed that represents _everything_ on my site, I can figure out how to use that to then handle sending Webmentions on the fly)
#
Zegnat [jacky]: I had to dig through my mind for a bit, but Datasette/Dogsheep comes to mind for me. Simon spoike about it last year https://indieweb.org/2020/East/personaldata. Some links there.
#
Zegnat The idea was to have a sqlite backend containing a mix of all sorts of data, and then putting a renderer/search tool on top of it for when you weren’t pulling at the raw data.
#
Zegnat Sounds similar to what you have to me, in that Koybe (I imagine) could end up storing all kinds of mf2 JSON from micropub, expressing all sorts of things.
#
Zegnat And if I completely missed the point, well, no harm in linking Datasette again, as I find there to be a lot of great inspiration in there :D https://simonwillison.net/2020/Nov/14/personal-data-warehouses/
[schmarty] joined the channel
#
[schmarty] that reminds me i've been thinking about https://phiresky.github.io/blog/2021/hosting-sqlite-databases-on-github-pages/ a lot lately.
#
aaronpk that was 🤯
#
[jacky] just came across https://github.com/rixx/goodreads-to-sqlite from that page, Zegnat. Definitely handy
#
Loqi [rixx] goodreads-to-sqlite: Export your (or other people's) Goodreads data to SQLite
#
Zegnat sqlite++
#
Loqi sqlite has 1 karma in this channel over the last year (2 in all channels)
#
[jacky] So my thinking is (hopefully) having a website that considers Micropub or its contents to be yet another 'data' source - it doesn't have the concrete concept of what a blog would look like
#
[jacky] tbh I might have to tinker on this more but my hope is to make the plumbing bits of the IndieWeb completely isolated from my site so I could even consider swapping them with no downtime
#
[jacky] (or little downtime lol)
[aciccarello], michael, samwilson and [jgmac1106] joined the channel
#
[jgmac1106] [snarfed] question fromt he microblog world. any open tix about accepting an RSS feed and spitting out a JSON Feed: https://micro.blog/manton/11516087
#
Loqi [manton] @DoctorMac Granary can output JSON Feed but doesn't look like it accepts RSS.
#
aaronpk ?
#
aaronpk there's a giant RSS button on granary.io
#
aaronpk oh that is only for the output format! odd
#
[tantek] consuming RSS is MUCH harder than publishing it
[KevinMarks] joined the channel
#
[KevinMarks] unmung consumes it
#
[jgmac1106] ahh thx tantek and I forgot about unmung
#
[jgmac1106] so they could go unmung, to hfeed, to granary and json feed
#
[jgmac1106] I mean as tantek note may look pretty ugly but filters and categories on microblog pretty solid
#
[KevinMarks] if you say input is atom does it understand RSS too? I thought granary used universal feed parser