#GWGsnarfed: The faces in a facepile. Are they supposed to link to the url or the author url? I remember some discussion about this, but I can't remember what was said.
#mblaneygRegorLove were you still logged in when you clicked that link? The idea with that comment is that it appears in my feed (since I'm the recipient) and then I can click the link to create the reply.
#mblaneyso outside the feed it loses it's context, will have a think about how to make that clearer.
#mblaneyyeah I should word that better. I don't want it to say "paid" because if I now pay you back $5, saying that you've only paid me $5 wouldn't make sense either.
eli_oat joined the channel
#mblaneymaybe just "You have a balance of -$10.00" instead?
#skippyI want to post retweets on my site, but want to preserve line breaks and formatting. So I *think* I want <blockquote class="p-content><pre>Some tweet content here</pre></blockquote>. I have CSS declarations for each of pre and blockquote, so the previous example seems to be getting the padding of BOTH of those.
#skippyinteresting! thanks for the pointer. Mostly, I don't want to (overly) mangle some of the more clever tweets people post involving specific formatting...
#mblaneyskippy you can also create a more specific selector: "blockquote > pre" and set the padding to 0 for just that case.
#skippymblaney: I tried that, and it never seemed to work... Maybe I was doing it wrong.
#gRegorLovemblaney: Does it make sense to just display separate totals of debits and credits instead of a balance?
#mblaneydo you have a page you can link to where you've tried it?
#mblaneyI added it as the last line to the inline style sheet, just: blockquote > pre { padding: 0 }
#skippywhen i try that, it still seems to be inheriting one of the other padding declarations. If I remove blockquote, but leave blcokquote>pre, then I get zero padding.
#mblaneyit still gets the padding from the blockquote, but it shouldn't be double padded. what browser are you using?
#Loqimblaney has 3 karma in this channel (24 overall)
#mblaneyno worries, strange just tried it on a mac too and it worked ok.
[jgmac1106] joined the channel
#jmacAbout how often does webmention.rocks regenerate its test pages?
#jmac(I am tenuously flinging some bespoke mentions at it now.)
#jmacWhoops, got it. Forgot I had my comments-hiding browser extension on.
#jmacWhee, Web::Mention passes all .rocks new-post tests. 💪
chrisaldrich joined the channel
#oodanii'm curious: micropub uses essentially one endpoint, with the request body deciding whether it creates, updates, or deletes. was a more restful design considered? something like POST /micropub/entries, PATCH /micropub/entries/5, DELETE /micropub/entries/32
#oodanithat of course only works for h-entry, but that's no showstopper: GET /micropub could return something like {"h-entry": "/micropub/entries", "h-event": "/micropub/events"} to advertise both the supported microformats and where they're located
#jackyoodani: one could make a wrapper but the micropub endpoints/spec is meant to be more RPC-friendly
#aaronpkWhy the sudden interest in the REST pattern?
#jackyand imo, REST wouldn't be efficient at scale for something like micropub
[eddie] joined the channel
#mblaneygRegorLove moving here from meta: thanks again! looks like not having categories published in your post has stopped it from completing the process though...
#gRegorLoveAh, I thought your site might salmention it to indiepay
#mblaneythis is all a bit goldberg-esque, but the flow is: if my site picks up the in-reply-to it will publish it as a comment, when it publishes a comment it salmentions the original links, then indiepay.me picks up on the whole thing.... so without the markup it doesn't make it!
#mblaneyI appreciate your feedback though and I'm busy fixing things :-)
#aaronpkThere was a good talk about api design a few months ago, I wonder if the video is online
#gRegorLoveso indiepay verifies both posts link back to it? that makes sense.
#mblaneyyour reply doesn't have to link to indiepay.me but it indiepay.me will check that it's in-reply-to the original post (and that you as the author of the reply are person-tagged in the original post)
snarfed joined the channel
#[eddie]Sweet! I’ll have to watch that video, looks interesting!
tomasparks, wagle, snarfed, KartikPrabhu, wagle_, loicm, jeremycherfas, swentel, tglobe and [kevinmarks] joined the channel
#Zegnat!tell GWG re facepiles and where to link them, have a look at sebsel’s implementation: https://indieweb.org/facepile#Sebastiaan_Andeweg it shows author portraits as links to people, and then reply-type-icons as links to posts.
KartikPrabhu, eli_oat[m], ancarda_, AngeloGladding and jeremycherfas joined the channel
#sknebelheh, at IWC Düsseldorf someone also asked about why Micropub is like it is. I guess if you use a "full-featured" web framework like Ruby on Rails or Django the REST-y approach is fairly natural to implement, and the one-endpoint design actually more confusing to implement., vs in other environments the one endpoint is easier in various ways
loicm, plindner and [kevinmarks] joined the channel
#[kevinmarks]also PUT and PATCH are very hard to implement well
#[kevinmarks]It needs you to replace the whole resource, which means you should read it all, change it and write it back. Now you have both a "did I parse it right?" problem and a possible race condition.
#[kevinmarks]It puts the correctness on the client, whereas using POST to modify makes it clearly the servers job
#ZegnatBut PUT always overwrites the whole thing right? So no parsing should be neccessary? Surely taking the HTTP request body and writing it to a file wholesale is actually less error prone than POST where you have to take the request body and parse it?
myfreeweb, schmarty, zoglesby, eli_oat[m], grantcodes and mindB joined the channel
#[kevinmarks]If you're just modifying a file, sure. If it's a structure that will be interpreted as something else, much trickier.
#[kevinmarks]It's superficially easier for the server, but puts more onus on the clients
#ZegnatIt is just as hard for a server to interpret structure send as the body of an HTTP POST as it is to interpret the exact same structure in the body of an HTTP PUT. I do not see a technical difference there in the slightest.
#ZegnatThe only question is, what does the PUT method mean and what sets it apart from the POST method. And that is where the RESTful design people come in ;)
#Zegnat!tell dgold should a PR for nanopub target the develop or master branch?
#ZegnatNot sure I follow. PHP’s server only knows what to do with a very specific POST request. It will try and parse a form-encoded HTTP request body on a POST request and give it to the user in a specified variable. Any other POST request it doesn’t do anything with it, just like with any PUT request.
#ZegnatThat list is cool [kevinmarks]! I wish they had actually written names in the name column though.
#ZegnatI happen to know that ISO 8601:2004 is for datetimes, but that is a lousy name
#jgmac1106so [kimberlyhirsh] sent me a link that I was able to follow here: http://schema.org/Course been getting a lot of push back from folks saying schema.org rdf and JSON-LD is the future and nobody will ever mark up their courses to make microformats2
#jgmac1106I disagree their big argument is always, “Well that is what Google wants.” I kind of see that as an exact reason to choose an alternative with greater control of the content creators.
#ZegnatStick to the format you like. But is anyone actually authoring JSON-LD in the edu circles?
#jgmac1106The other push back I have been getting is any standardizartion must be done with large standards boards and the major edtech vendors .
#ZegnatWell, if there is a big presedent for using an already documented vocabulary in JSON-LD, I am not super sure what the argument is for copying it to microformats?
#ZegnatThere are arguments for why to use one data format over the other, if you are starting out. But if there is already an existing format that people agree on, why introduce a competing one? That’s an important question to answer.
#jgmac1106thx @zegnat feels like a lot of stuff just about to fall in place, everything in a bit of transition from weird xml implementation
eli_oat joined the channel
#ZegnatThat said, I do not know a whole lot of JSON-LD parsers outside of jkphl’s. So not sure anyone is consuming what is being published?
#jgmac1106part of my reason is where the control ends up falling. If those in higher ed are going to control what we publish and how we publish keeping things as simple as possible is better in the long run. Common markup hard to implement stuck behind proprietary software leaves us still beholden to the publishers and platforms IMHO
#jgmac1106Looking at the schema.org spec for example their are more classes for protecting copyright and institutions than their are for the actual learning the object contained. I guess thats, “My Why?” but I am wondering if I am trying to create a groundswell for a battle that may have been over for some time
#ZegnatYes. If you are talking to folks who already know how to author HTML (or even already publish it!) it might be a lot easier to get them to add microformats than JSON-LD (which will be a choir to author by hand, IMHO).
#ZegnatBut if the people you are talking with are going to use specialised tools anyway. Whether those tools output HTML+Schema or HTML+mf2 isn’t really an issue, right?
#jgmac1106[zegnat] probably right. Need to remember the indie in indieweb and that this might be a small niche who reject the large corporate learning solutions
#ZegnatBut if the “large corporate” solution is Schema Course, why do you need to go against their markup at all? I guess that’s my main question
#skippyis there a succinct explanation for the benefit of microformats like "h-cite" over their HTML equivalent (<cite>) ?
#Zegnath-cite allows for marking up sub properties where <cite> is just “the title of a work”
#ZegnatYou could actually mark-up the p-name of h-cite with a <cite> element. But thanks to the multiple properties of h-cite, you can also add a p-author outside of that <cite> and it is still part of the same citation.
#skippykind of. thanks. the h-cite wiki entry says "If you are displaying the entire entry, then it does not need an h-cite." which confuses me a bit on how to display reposts...
#LoqiA repost on the indieweb is a post that is purely a 100% re-publication of another post. The act of reposting is an umbrella term that covers the general practice of republishing another post typically on the same service or silo, but more and more across sites https://indieweb.org/repost
#ZegnatThough you may want e-content instead of content on the h-cite?
#skippyi thought i was told that i should use e-content only on my original contnet, and not on reposts.
#jgmac1106[skippy] my succint reason is because I can. It is easier and more useable for an individual teacher. I could quickly throw up a template site and get people to remix it. I don’t knwo how to do that with JSON-LD. I guess that is my reason. And isn’t the phrase “large corporate’ enough motivation to go against?
#Zegnatskippy? Oh? I mean, it is up to you. I would use e-content for reposts. Because if I repost something with an image or links in it, I want that to be kept
#jgmac1106oops meant [zegnat] but thanks for helping me think this through. Lot of people saying I am chasing windmills right now
#skippyi havent fully internalized MF2 semantics, so I dont fully grok the difference between p- and e- stuff.
#Zegnatp- = plain text, e- = embedded HTML. So p-* extract only the plain text value of the element you put it on, while e-* will also return a copy of the verbatim HTML markup
#Zegnatjgmac1106, if the individual teachers you are targetting are comfortable authoring HTML, it makes a lot of sense to give them remixable HTML snippets, yeah.
#skippyok... so for a repost, e-contnet might make sense. but then on a reply, in which i blockquote the post to which i am replying, does it make sense to have two e-contents?
#ZegnatThe reply context blockquote can use the exact same mark-up as the repost. It can be an h-cite within a u-in-reply-to. So completely fine to use e-content on that h-cite, and have a separate e-content on your h-entry
#[kevinmarks][jgmac1106] looking at the Course schema, there is a case there for a microformat with h-events for the instances. The value you cold potentially offer is RSVPs
#[jgmac1106][kevinmarks] I could see RSVPs being a method for taking attendance
#[kevinmarks]schema has a way of marking those up, but not a mechanism
#skippywoo! Finally wrote a (more) decent README for my micropub endpoint!
#Loqiskippy has 4 karma in this channel (9 overall)
#jeremycherfasSkippy++ zegnat and I have been working our way through getting it to work for me. And we're getting there, slowly, although I won't be able to do more until next tuesday at the earliest.
#skippyadded a new blurb about my own setup, to give a little more clarity (I hope)
snarfed joined the channel
#skippyaaronpk: thanks for sharing that API video. Interesting stuff.
loicm and leg joined the channel
#Zegnatdgold, the promised PR is finally sent. Sorry for the slow down. It got through all my testing but obviously you should also test before merging.
#Zegnatjeremycherfas was testing your endpoint and had to do some code changes to get it to work with a self-hosted token endpoint. (Mintoken in that case.)
#ZegnatThe function still uses the $configs variable as a global, dgold. Because I wanted to keep my changes the least invasive. So no changes to the function signature
#ZegnatProbably doing the Mintoken security update this weekend, then I only have issues about error reporting left to tackle. And those shouldn’t be breaking.
#dgoldthat's best, I tried changing that around to not use the config as a global, but i hit the Dunning-Kruger Boundary
#jeremycherfasAnd I'll be ready to stress test to the max next week, probably Tuesday.
barpthewire joined the channel
#skippyi havent looked closely at the mintoken code yet. what was the driver for SQLite storage? Ease of development?
#skippyfor myself, I see only a small number of authenticated clients connecting to my micropub target. Quill, Omnibear, Indigenous, Android Micropub. DB seems a littel overkill; but I'm quite sure I'm missing somethign fundamentally important.
snarfed joined the channel
#ZegnatOne Mintoken instance can be used by multiple sites, and as any site can have any number of tokens, this just seemed easiest.
#dgoldnot like the memory hogs of the larger databases
#ZegnatMore importantly, you don't need separate backup processes for it.
#skippyyup. i'm super familiar with SQLite; and generally a big fan of it.
#aaronpki'm usually not a fan of sqlite for web apps, but this does seem like a good use of it
#skippylooks like I forgot to add SQLite to my PHP container. *sigh*
loicm joined the channel
#ZegnatMintoken could be tweaked to accept any PDO DSN, so people can pick their own DB, I guess. But I don’t see myself changing it to a non-db system
#skippyPHP Notice: Use of undefined constant SODIUM_BASE64_VARIANT_URLSAFE_NO_PADDING - assumed 'SODIUM_BASE64_VARIANT_URLSAFE_NO_PADDING' in Command line code on line 1
[chrisaldrich], snarfed and [schmarty] joined the channel
#skippy"Based on the German legal literature I have assessed, this would make the GDPR applicable to a personal website as soon as its content has any connection to professional or commercial activity, hence not only applying to freelancers, who undoubtedly are affected, but e.g. to any web professional who discusses web technology on their personal site "
#Loqi[Daniel Goldsmith] Sebastian, first of all, thank you for your detailed write up on this issue. I think much of your roadmap is worthwhile, and of great interest.
I cannot, however, say that I am convinced by your contentions regarding the effect of GDPR and indieweb ...
#LoqiIt looks like we don't have a page for "indiepay.me" yet. Would you like to create it? (Or just say "indiepay.me is ____", a sentence describing the term)