#KevinMarksUX improvement suggestion for indiauth.com in the email message put the code in the beginning of the subject so I can read it from the notification without switching apps
#jackyI've seen a few services do that (one bank I used to bank with)
eli_oat1, [jgmac1106], snarfed1, [manton], [davidmead] and [eddie] joined the channel
#petermolnar!tell cleverdevil I tried to use ditchbook on a current facebook page export, but it doesn't like it, due to mismatches in available/unavailable dictionary elements, eg. status without media elements.
#Zegnatjacky I thought the indieauth spec also explicitly mentioned the query params. Hmmm...
iasai, barpthewire and swentel joined the channel
#swentel!tell [eddie] can you pull on the indigenous website. Added a link to the f-droid download page. Thanks!
#[Vincent]I’ve noticed that when I use social readers the static images from mapbox.io dont render correctly. Not sure if that’s some weird specific issue to me or something you need to bear in mind.
#[kevinmarks]You can do dynamic maps with leaflet and openstreetmap tiles, but not sure about static.
#LoqiWhile self-hosting is comparatively easy for content formats like photos, same does not apply to interactive maps – given the sheer amount of data and/or the complexity of its processing. Experimenting with OpenMapTiles, it for the first time appea...
#sknebel[schmarty]: ohh, good point, he linked such scripts already
#[Vincent]You are free to copy, distribute, transmit and adapt our data, as long as you credit OpenStreetMap and its contributors. If you alter or build upon our data, you may distribute the result only under the same licence.
#GWGThey don't have a simple static map generator though
#sknebel[Vincent]: it's a bit more complicated than that. the map data is one thing. another thing is the pixels being generated from it, and the terms under which you can ue APIs giving you the pixels is yet another
#[Vincent]ah so you need to do something messy to translate it in to a jpg?
#sknebelvarious places give you tile images, but often free tiers are limited. and GWG ideally wouldn't want tiles, but larger images already combined he can store locally
#sknebelthe staticmaplite script above could be interesting to assemble tiles yourself, then the issue is "only" finding a set of tile providers that work
#aaronpkgenerally you'll want to find a way to use a proper geospatial type in the database, but that isn't going to work on the lowest common denominator for wordpress
#GWGaaronpk, that's the question. Usually the WordPress meta database, but I might need a custom table
#GWGWordPress does allow for custom tables, but I always avoid them
#aaronpkyou'll definitely want a custom table at least
#aaronpki think if you want to avoid the actual geospatial extensions your best bet is to use an indexed column that is an integer that is the geohash https://gis.stackexchange.com/a/92331
#sknebeltrying to find a small-ish setup to use OSM venues is still on my todo list, but different priorities right now before diving into something like that
#aaronpkyes also you're likely not going to encounter an issue with a full table scan until you end up with several thousand rows anyway
#aaronpkactually probably not even til tens of thousands
#GWGI am not limiting it. It was back to implementation.
#ZegnatEspecially if waiting rooms are involved where they still want to use their device without publicly broadcasting whos waiting room they are in, he
#GWGRight now, it is an array of name, lat, lng, and radius and it goes through the array item by item and stops when it finds one that matches
#GWGIf there are more than a few, it could slow things
#aaronpkeh don't worry about it, that math is cheap
#aaronpkit'll only be a problem if there are thousands of items in the array or if that code is run many times a second
#GWGSince I store location in each post, I probably should store the result in the post.
#aaronpkyes, i would store whether or not location should be displayed in the post itself rather than computing it on each pageview
#ZegnatThat is interesting though, do you expect the geofence to apply retroactively? Because then you are looking at checking every post when the list of fences is edited.
#ZegnatThat might be way more expensive for someone like chrisaldrich than the lookup for every pageview?
#aaronpktake the lazy approach, and give the user a checkbox when editing the post so they can turn it off and on per post, and only set the checkbox automatically for new posts
#GWGIt is currently set by a default option, if not by the user
#GWGAnd Indigenous allows it to be set by Micropub..thanks swentel.
[tantek] joined the channel
#GWGSo, this is having it set by default if the location is in a radius around one of the stored zones
#GWGAnd he just added an input to send a location name as well.
#ZegnatOh that makes sense, have only the default value depend on excluded zones, and never touch already existing/published posts
#aaronpksounds like you have a good mechanism in place already. i'd make the new zone thing just set that default value if it's not provided by the user or in the micropub request
#GWGZegnat, exactly. But by storing that it was found to be in a zone, if the zone changes, it could update
#ZegnatTheoretically you should be able to merge h-cards as long as they share their url value, I think, GWG. So rather than finding 1 just concat them together?
[davidmead] joined the channel
#GWGZegnat, I am trying to normalize feeds for Microsub.
#ZegnatOh, sorry, I was thinking you were only talking about deciding which h-card to parse
#GWGI am trying to make it when I read in Indigenous or Together that I always have a name or photo
#Zegnatyeah, so I am saying rather than having to create criteria for finding the right h-card, just merge them together so you only ever “find one” :) Though I am not sure anyone is doing it that way today
#ZegnatTheoretically, because the url property on the h-card identifies it, and other h-cards with the same url property can be merged into it. So someone can have an h-card on their h-entry, h-feed, and outside in their footer, yet those could probably all be merged into just one card.
#ZegnatMaybe it doesn’t help, maybe it does, was just a random thought I had
#GWGI anticipated a top level h-card, but I assumed that if you had one, you wouldn't also have it on the h-feed and on every individual post. I ended up storing all h-cards in a temporary array and merging them together, then adding where missing
tbbrown joined the channel
#GWGI wonder if this weekend, I can find out how to return h-cards in Indieauth responses
snarfed and chrisaldrich joined the channel
#ZegnatI’d love to call in for that discussion. I think a lot of projects can be helped by that.
#GWGWell, we discussed it the other day, but no format decided
#GWGI am already returning the WordPress user id in there
#ZegnatYes, I read the discussion, and the bit about the profile scope (which I found very interesting)
#Loqi[snarfed] #350 facebook: v2.0 api no longer allows lookup by username
#[cleverdevil]ditchbook is a tough project, because in order to really make it work well, I need to have a large number of people's facebook exports to work with.
#Loqi[cleverdevil]: snarfed left you a message 5 days, 4 hours ago: hey podcasters, we don't yet have exact audio file size (in bytes) in mf2. how should we represent that? i'm looking at converting to RSS, where it's required. more: https://chat.indieweb.org/dev/2019-02-17#t1550418872983700
#Loqi[cleverdevil]: petermolnar left you a message 13 hours, 20 minutes ago: I tried to use ditchbook on a current facebook page export, but it doesn't like it, due to mismatches in available/unavailable dictionary elements, eg. status without media elements.
#snarfedjacky: this is what the micropub syndication param and query is for, right?
#aaronpkjacky: there's also the micropub overloading for syndication. Like how silo.pub handles the silo APIs and gives you an API to it that happens to look like micropub
#jackysnarfed: this could add more values to populate that but yeah!
#snarfedyup. micropub definitely seems like the right cross-tool protocol for POSSE/syndication going forward. bridgy using webmention was really just a historical accident because i built it before micropub.
#jackyLike if I wanted to just click in a UI to add "Twitter" support
#snarfedGWG: right but this is beyond that. you can outsource the targets to bridgy and still hard code or have the user enter their silo accounts in site setup
#snarfeddynamically discovering them from the service seems like overkill, esp since it's usually only a small set that very very rarely changes
#snarfedmanual until it hurts, and all. making it this dynamic and generic seems a bit like overengineering to me.
#GWGsnarfed: I built something in this area into Syndication Links. Someone can implement a class that tells it what the strings are to display, and a function that does the POSSEing, and it handles the Micropub target registration and checkboxes in the Classic Editor.
#GWGSo far, no one has actually written a target except me, but anyone could very quickly.
#jackysnarfed: that's fair. I'm looking a bit further ahead (towards users who might not have a deeper grasp into the indieweb of things but do want to try this out)
#[eddie]I agree the UI issues are for more relevant than the need for try to automate that stuff. Also aaronpk's site because it uses Micropub supports a kind of IndieAuth like "adding" if I remember correctly
#[eddie]While the Generations model isn't the best the one thing it DOES help with is the concept that we have to iterate. We can't build for Gen 4 until we've built for Gen 3
#[eddie]the UI and tools will progressively iterate until they reach general public. I don't believe we can jump from our current setup to general public
#[jgmac1106]...for my badges...instead of building them all one at a time, I could build them all at once and uncomment out as they are earned....just a stopgap until I build a proper system for webmention badges
#ZegnatOh interesting. Theoretically the Webmention receiver is not expected to accept URLs inside commented HTML as valid mention targets
#snarfedfor html it does recommend to look for link, img, or video, which is similar but narrower
#Zegnatsnarfed, I am extrapolating from the fact that a receiver should look for href/src attributes. Also from the GitHub issue I linked here. If you are checking HTML for mentions, you should be using an HTML parser, which will turn a comment into just a big string.
#[jgmac1106]just the first one, it looks closest to what I need yet
#ZegnatThough what exactly a link is, may still be up for debate ;) E.g. I initially wrote in that GitHub issue that <form action> could probably be ignored. But now that forms are getting rel attributes...
#ZegnatSo, yes, Webmention spec may be a bit vague on the subject. But for it to define what are and aren’t HTML links is probably super-duper out of scope :D
[schmarty] joined the channel
#sknebelIf I remember right from skimming it, bridgy fed might just stringmatch. But that's a bit of a special case
#snarfedyeah mine all blindly string match. but they abuse webmention as API trigger instead of actual mention, so ignore them.
#sknebel14 hours until sknebel, you have a table to update
#LoqiI added a countdown scheduled for 2019-02-23 1:51pm CET (#6494)
#[tantek]wait a minute, since forms are getting rel attributes, and we can use a form for webmention sending UI, does that mean it can double as discovery for the webmention endpoint too?!?
#Zegnat[tantek], yep, it was one of the few cases where it would actually make sense. We discussed it a little when annevk came by the #microformats channel
#ZegnatWonder if the mf2 spec needs changing or not. It mentions rel parsing needs to happen for tags that constitute a “hyperlink element”. While forms are now links, I don’t know if there is a spec-difference between HTML Links and HTML Hyperlinks. Or if mf2 spec hyperlink is even the same as the HTML spec hyperlink
#[grantcodes]I wouldn't have thought so tantek. I only remember seeing one and it had a different syntax to that
#[grantcodes]I think also my chatbot could be run on the command line but never bothered trying yet
#[tantek]so here's a thought, what if we moved the @indiewebcamp twitter to micro.blog, and then had it cross post from micro.blog, then have Loqi use micropub to post to micro.blog/indiewebcamp instead of posting to Twitter?