#dev 2018-07-30

2018-07-30 UTC
#
[jgmac1106] I was going to stick each image on every page into a Twitter card in header and did not know if I should add u-featured inside of h-entry.... Then I am trying to fork my cc flickr attribution tool to spit out correct mf2 cite of a photo
#
[jgmac1106] Two different random chains of thought... I should stick to philosophy
#
gRegorLove If that's a representative image for the post, then u-featured.
#
[jgmac1106] Thx gRegorLove it is
snarfed joined the channel
#
[jgmac1106] Wow I stink at media queries and CSS Grid... Have way bigger problems to solve than u-featured. Enjoy your night all
snarfed joined the channel
#
[jgmac1106] .... If only I didn't listen when someone said forget bootstrap learn css grid...
#
gRegorLove I found gridbyexample.com helpful and starting with smaller components vs. entire page layout (like the author section above each post on indiebookclub)
marylicious joined the channel
#
jgmac1106_ thx [gRegorLove] I did way better on my syllabus but knew more then: https://edu307syllabus18summer.jgregorymcverry.com/
#
Loqi EDU 307: Children's Literature and Literacy
#
jgmac1106_ and didn’t use any of the shortcuts, i think my course template will need a total overhaul, when I spell out the grid I do it right…the holy grail template I used…wasn’t ready for
#
tantek__ ok different question. what's the point of a Twitter card?
#
tantek__ or to be more blunt, what's the point doing any extra proprietary platform-specific work for an ever more toxic platform?
#
jgmac1106_ [tantek] to get an image to unfurl, cuz its pretty, and that is where the audience I would like to onboard to more indieweb approaches live. Be nice if a any link shared on Twitter with u-feature would post like a Twitter card
#
tantek__ do your tweets get the "engagement" you think they're getting to justify the work?
#
tantek__ trying to quantifiable ask whether "the audience I would like to onboard" is actually even seeing your tweets at all
#
tantek__ (FWIW I have found that tweet engagement numbers continue to crash, as people seem to be leaving or using Twitter less)
#
tantek__ and their user #s actually dropped last quarter
#
jgmac1106_ edutwitter is quite strong. I still attend 7-9 hourly chats a week during school year
#
jgmac1106_ but yeah they are in trouble….and if every single app needs its own developer account…so much for playing in that playground for Johnny Q. Public
#
tantek__ exactly
#
KartikPrabhu boop!
#
Loqi KartikPrabhu: tantek__ left you a message 7 hours, 24 minutes ago: re: https://chat.indieweb.org/dev/2018-07-29/1532881135924700 - that's inaccurate. There's no such thing as a "linkback" (it's a made-up umbrella term AFAIK, not a specific technology), and trackback is not "just a ping" - it sends the full payload (which is why it was even spammier).
renem joined the channel
#
@OAuth_io We will be talking about 'The Many Flavors of OAuth' at @APIdaysGlobal San Francisco about #oauth2 and briefly covering identity layers #openidconnect #oidc and #IndieAuth. We have a few tickets to giveaway. Please register with code 'Soonhin' at https://www.apidays.co/sanfrancisco. See you! (twitter.com/_/status/1023746803395514368)
[matpacker], chrisaldrich and mblaney joined the channel
#
gRegorLove jgmac1106: re: Twitter Cards, you could test with their validator https://cards-dev.twitter.com/validator and document minimum viable alternatives like we did on /OGP.
#
gRegorLove For example, in absence of OGP, Facebook will choose the first image that's at least 600px wide for the image when you share a link. Maybe Twitter does similar.
#
gRegorLove Can document on /Twitter_Cards#How_To
#
tantek__ perhaps we need an expansion of /Twitter_Cards#Why
#
gRegorLove Perhaps. /OGP doesn't have one either.
#
tantek__ yeah
#
tantek__ we have /link-preview for the general reason
#
tantek__ but perhaps those that care enough to publish og: or twittercards can help fill out the platform specific /OGP#Why and /Twitter_Cards#Why
#
tantek__ and maybe we can include a minority opinion of #Why_bother
#
gRegorLove heh
stevestreza, [kevinmarks], KevinMarks, [snarfed] and [miklb] joined the channel
#
[miklb] would this be an accurate statement? mf2 is the semantic markup in the theme that facilitates communicating the context of the content.
#
KartikPrabhu [miklb]: depends on the context of the usage
#
[miklb] I’m trying to gather all of my thoughts from today and put it into something I can publish.
#
[miklb] https://gist.github.com/miklb/5810035b8a4414dbf88293e397acfddb
#
[miklb] that might be as far as it gets. 🤷‍♂️
[pfefferle] joined the channel
#
KartikPrabhu [miklb]: couple of things about that post. microformats is not mandatory to using webmentions. one can use mf1 classes and mf2 at the same time. So WP could add mf2 classes to their markup without removing the mf1 classes
#
KartikPrabhu so it i entirely doable without breaking anything
#
jeremych- [miklb]: Good summary.
#
[miklb] really. glad you solved that.
#
KartikPrabhu and lastly, WP should actively discourage people from using mf classes for styling
#
[miklb] we can all go home now.
#
KartikPrabhu I am home so :thumbsup:
leg1 joined the channel
#
[pfefferle] [miklb] thanks for the summary, but WordPress core support of mf2 is not the solution of all of our problems... core only sets a few of the needed classes, like h-feed and h-entry... but theme authors still have to set p-name and e-content... so the core support might be worse than the current state, if theme developers will not catch up with implement mf2 support... In the end it is still up to the theme developer...
[matpacker] joined the channel
#
dougbeal I don't understand the types of blocks in Guetenberg very well, but is that were mf2 support would really win?
#
[pfefferle] Might be a chance, but still hackt
#
[pfefferle] hacky
#
dougbeal Is it? Wouldn't that be where you would know by block type the mf2 markup it should get?
#
[pfefferle] [tantek] for a WordPress user it is easier to add OG, because it is theme independent
#
[pfefferle] [dougbeal] but you still have the problem to mark up the e-content container
#
[pfefferle] Or the header
#
[pfefferle] Or the author markup
ben_thatmustbeme and [kevinmarks] joined the channel
#
[kevinmarks] Blocks seem more microformats friendly in that they are chunks of meaning, but I haven't looked at details
cweiske joined the channel
#
tantek__ [pfefferle]: that is why keep coming back to specifying /metaformats explicitly, so we can backcompat (re)define OG et al as their microformats equivalents for the page
#
tantek__ and frankly to be able to set the root microformat class name on the <html> (or <head>) element, and then the rest in <title class="p-name"> and <meta> tags
#
tantek__ which would be the right thing to define first, and then define how to process og: and other meta props as backcompat for mf2 class names
jeremycherfas, KevinMarks, TripFandango and [matpacker] joined the channel
#
@PraveenPal4232 Webmentions: Enabling Better Communication on the Internet #web #webdev #webdevelopment #webdeveloper #Internet #html #css #javascript #reactjs #reactnative #AngularJS #vuejs #nodejs #frontend #UserExperience #userinterface #WordPres #cms #webapi https://goo.gl/kazQJC (twitter.com/_/status/1023841883531825152)
KevinMarks, [jgmac1106] and jgmac1106 joined the channel
#
@ProtoSyte Webmentions: Enabling Better Communication on the Internet · An A List Apart Article https://alistapart.com/article/webmentions-enabling-better-communication-on-the-internet (twitter.com/_/status/1023910946396413953)
eli_oat joined the channel
#
tantek__ !tell KartikPrabhu I think saying "microformats is not mandatory to using webmentions" while hypothetically true, is not true for practical purposes. For example have you yourself implemented some way of using webmentions without microformats? If you have, then great, document that and provide it as an example. If you haven't then better to explicitly provide that "theoretically" or "according to the spec" caveat
#
Loqi Ok, I'll tell them that when I see them next
#
jgmac1106 [tantek] will work on Twitter card page, but I don’t really care about 3-5x engagement with image, It’s pretty, it’s about storytelling with a nonlinear but continous design narrative, and about having fun
#
skippy good day.
#
jgmac1106 .oohh and u-featured didn’t work syndicated or in native twitter app, but that wasn’t from my domain, doing that now
#
jgmac1106 hey Skippy
#
@Brilliantcrank Webmentions sounds awesome. I wish we had had this back in 2002, when everyone was writing. https://alistapart.com/article/webmentions-enabling-better-communication-on-the-internet (twitter.com/_/status/1023923807004897280)
#
martymcgui.re edited /podcast (+182) "/* Marty McGuire */ add deets about h-feed on TWITIWAE" (view diff)
snarfed, [snarfed] and [tantek] joined the channel
#
www.svenknebel.de edited /IndieWebCamps (+73) "/* 2018 */ add SF" (view diff)
#
www.svenknebel.de edited /IndieWebCamps (+110) "/* 2018 */ add Berlin, Düsseldorf" (view diff)
chrisaldrich and [grantcodes] joined the channel
#
[grantcodes] RE: WordPress microformats: I doubt it is a good idea, but definitely more possible in the wp ecosystem would be to have a plugin output hidden `data` tags with all the required markup. It's dirty, but it would work with every theme
#
Loqi [grantcodes]: jgmac1106 left you a message 6 days, 4 hours ago: did you fix your webring, my fibonacci found poem still stuck on your site?
[miklb] and stevestreza joined the channel
#
boffosocko.com edited /export (+389) "chrisaldrich & cleverdevil examples" (view diff)
#
aaronpk lol [grantcodes] that's an interesting idea
#
jgmac1106 [aaronpk] my poem is still stuck….[grantcodes] only has one job right now
#
gRegorLove jgmac1106, I'm not seeing media queries in your glitch. Did you take them out?
#
jgmac1106 yes, I was able to do them perfect for auto rows with set widths, but as soon as I started playing in grid and had some boxes span mutiple columns I kept messing up
#
jgmac1106 I never predefined any grid areas but thought I could do a media query for wrapper and set the grid columns to 1, that didn’t work (or I typed it wrong) then tried to fo a media query and give set each row to span 1-8
#
jgmac1106 didn’t work
[ryan339] joined the channel
#
[grantcodes] Ohh sorry jgmac1106
#
[grantcodes] Would it work if I have the links but they are hidden? Never figured out how to include them without mangling my design
snarfed joined the channel
#
jgmac1106 no apologies, still et credit for the first indieweb ringbreaker….well I did build you this: https://wonderful-clave.glitch.me/ not pretty enough?
#
jgmac1106 why not just the right and left arrow in the same icon set as your social share buttons?
#
gRegorLove jgmac1106: It appears to work if I change grid-template-columns on the wrapper and turn off the grid CSS for the three boxes (just in local dev tools)
#
gRegorLove That makes the three boxes stack
#
jgmac1106 gRegorLove++
#
Loqi gregorlove has 50 karma in this channel (266 overall)
#
jgmac1106 turn off?
#
jgmac1106 [grantcodes] and I am of course kidding, there is no requirement to add the webring, I can start a new found poem…Loqi just made me laugh
#
gRegorLove jgmac1106: I disabled the properties on the 3 boxes in dev tools, so "default". Looks like it also works changing the grid-column values to 1.
#
jgmac1106 thx, will try
#
gRegorLove As long as the parent element's grid-template-columns is set accordingly
#
jgmac1106 yes what did you change this to? grid-template-columns
#
gRegorLove grid-template-columns: auto;
#
jgmac1106 and you could still have box A span 1-6 and box b span 6-8 on desktop?
#
gRegorLove Yeah, with media queries you can change those properties
#
jgmac1106 edventaullay what I am going to do is make a feed for all their comments and all the class notes
#
jgmac1106 thx
#
gRegorLove "ed" ventually. I get it, haha :)
#
gRegorLove hashtag academicpuns
#
jgmac1106 ha ha
#
snarfed hey sknebel just fyi, i plan to work on merging mf2py PRs and cutting a release tomorrow during IWC SF hack day, roughly 16:00-23:00 UTC, if you're interested in working together
#
sknebel sure!
#
snarfed great!
#
snarfed oh also sknebel mind if i close https://github.com/snarfed/bridgy/issues/837 ? or do you have more thoughts?
#
Loqi [sknebel] #837 twitter publish: posts unsupported markdown(?) formatting
#
sknebel can be closed
#
snarfed thx
#
gRegorLove snarfed, if you get a chance, run it against microformats/tests. There's been a few updates in the last week.
#
sknebel will do
[chrisaldrich], jackjamieson and [cleverdevil] joined the channel
#
KartikPrabhu hi Loqi
#
Loqi KartikPrabhu: tantek__ left you a message 4 hours, 12 minutes ago: I think saying "microformats is not mandatory to using webmentions" while hypothetically true, is not true for practical purposes. For example have you yourself implemented some way of using webmentions without microformats? If you have, then great, document that and provide it as an example. If you haven't then better to explicitly provide that "theoretically" or "according to the spec" caveat
[kevinmarks] joined the channel
#
[kevinmarks] I have code to run them as tests in mf2py, but lots fail. I'll see if I can get them closer. I wonder if we should segment the tests differently, and add some of the newer, more focused less synthetic ones.
#
[kevinmarks] That could be my hack day contribution
#
[cleverdevil] I looked into performance issues for the Micro.blog focused "On This Day" service I built, and pretty much all of the slowness is the initial parse of large archive pages using mf2py.
leg1 joined the channel
#
[kevinmarks] OK. Can you profile that further inside mf2py?
#
[kevinmarks] Snared reported some recent slowdowns too.
#
[cleverdevil] Yeah, I need to figure out what parser is being selected.
#
[cleverdevil] I am deploying to AWS Lambda, so I don't have control over the system, really.
#
[cleverdevil] I am specifying lxml
#
[cleverdevil] Even on my laptop, though, its slow.
#
[cleverdevil] To be fair, the page is *enormous*.
#
[kevinmarks] That's a good case for profiling then.
#
[cleverdevil] This is the page I am parsing -> http://cleverangel.org/archive
snarfed joined the channel
#
[kevinmarks] Between beautiful Soup, Html5lib and lxml there's a lot of places that may be copying strings or something.
#
[cleverdevil] Indeed.
snarfed and KevinMarks joined the channel
#
aaronpk content negotiation #fail... I forgot that I have a cache on my web server and it's not taking the content type into account when deciding to return from the cache
snarfed and [snarfed] joined the channel
#
[snarfed] conneg--
#
Loqi conneg has -1 karma in this channel (1 overall)
KevinMarks joined the channel
#
aaronpk welp
snarfed, tantek__, [schmarty] and [miklb] joined the channel
#
aaronpk starts down this nginx caching rabbit hole
#
aaronpk lol someone's markdown parser is a little aggressive https://www.digitalocean.com/community/tutorials/how-to-setup-fastcgi-caching-with-nginx-on-your-vps
#
aaronpk omg how do cache work
snarfed joined the channel
#
gRegorLove haha
#
KartikPrabhu :magic:
#
gRegorLove https://imgflip.com/i/2ez4fg
renem joined the channel
#
sknebel aaronpk: what are you trying to do?
#
aaronpk I need to somehow take into account the content type or accept header in the cache key
#
aaronpk so that I can cache the different representations of the URL separately
#
aaronpk since right now they are clobbering each other
#
aaronpk maybe I can just include $http_accept in the cache key? not sure browsers all send the same exact accept header t ho
#
aaronpk well that did *something*
#
aaronpk good enough for now
#
aaronparecki.com edited /conneg (+281) (view diff)
#
tantek.com created /messages (+23) "r" (view diff)
#
aaronpk closes that rabbit hole
#
snarfed aaronpk++
#
Loqi aaronpk has 167 karma in this channel (1691 overall)
#
snarfed conneg--
#
Loqi conneg has -2 karma in this channel (0 overall)
#
tantek__ is anyone doing anything with "shopping" on their own site?
#
aaronpk I have a few amazon wishlists that I keep thinking should be on my site instead
#
aaronpk closest i've done on my site I think is my page of product links/mini reviews
#
tantek__ that's related, but after the fact. the shopping has already happened
#
tantek__ we don't seem to have much besides /pay and /want
#
tantek__ doesn't rhiaro keep track of all her acquisitions / disposals of stuff somehow?
#
aaronpk yeah she has posts for everything she bought
#
snarfed venmo style
#
dougbeal.com edited /next-hwc (+0) "Redirected page to [[events/2018-08-08-homebrew-website-club]]" (view diff)
Toffee and [pfefferle] joined the channel
#
tantek__ what is purchase
#
Loqi An acquisition within the IndieWeb is a type of post that enumerates purchases, gifts, donations, (stolen?!) and found things potentially with additional information like photos, descriptions, specifications, price paid, date/time acquired, location acquired from, and condition https://indieweb.org/purchase
#
tantek__ close enough
#
tantek__ what is shopping
#
Loqi It looks like we don't have a page for "shopping" yet. Would you like to create it? (Or just say "shopping is ____", a sentence describing the term)
#
tantek__ shopping is /acquisition
#
Loqi ok
#
loqi.me created /shopping (+24) "prompted by tantek__ and redirect added by tantek__" (view diff)
[miklb], benwerd, benwerd_, [pfefferle] and tantek joined the channel
#
@jgmac1106 Getting my first webmentions from students blogs. It is a good reminder with Facebook shutting down their API we don’t need the silos. Everybody just needs a #DoOO and we can #IndieWeb (http://jgregorymcverry.com/5316-2/) (twitter.com/_/status/1024028483679346690)
snarfed and tantek__ joined the channel
#
tantek__ there's an abuse I've seen of the proprietary meta markup of placing text and images in them that is nowhere visible on the actual page
#
tantek__ e.g. in https://www.theregister.co.uk/2016/02/18/remember_wordpress_pingbacks_the_w3c_wants_us_to_use_them_across_the_whole_web/
#
aaronparecki.com edited /rel-alternate (+1001) (view diff)
#
tantek__ has: <meta property="og:image" content="https://regmedia.co.uk/2015/01/07/tommy_lee_jones1.jpg?x=1200&y=794"/>
#
tantek__ and in a jsonld data island: "image":{"@type": "ImageObject","url":"https://regmedia.co.uk/2015/01/07/tommy_lee_jones1.jpg?x=1200&amp;y=794","width":1200,"height":794},
#
tantek__ but of course, nowhere in the article is the image of Tommy Lee Jones looking skeptical
#
aaronpk I remember that
#
tantek__ anyone else seen abuse like that? where the /link-preview shows some rando and/or baity image, which is not present in the article, or shows a clickbaity title which is not on the article itself?
#
aaronpk it's the preview image when the article is shown in a list view https://media.aaronpk.com/Screen-Shot-2018-07-30-13-32-31-GpAwWesmw7.jpg
#
tantek__ this is one of the problems with invisible metacrap - it enables and encourages abuse of this sort
#
tantek__ but that's deceiving
#
aaronpk yeah it's not super representative
#
tantek__ we don't know that unless we document it
#
tantek__ also - they clearly want to be abusive on Twitter: <meta name="twitter:card" content="summary_large_image">
#
aaronpk I mean it's not representative of the post permalink
#
tantek__ make the fake / deceptive image LARGE on Tweets!
#
tantek__ ah yes
#
tantek__ could make for an ironic poke at The Register
[snarfed] joined the channel
#
[snarfed] http://www.kevinmarks.com/partialsilos.html
#
tantek__ The Register knowingly abuses defacto web standards OGP and Twitter Cards as well as uses the image of Tommy Lee Jones (presumably without his permission) to promote their article(s)
#
Loqi Decaying Silos as dead malls 2018-01-20
#
tantek__ "While The Register chicken-littles about theoretical web standard Webmention link spam they have never seen, they do so with a real world abuse of de facto web standards OGP and Twitter Cards to cause Facebook, Twitter, and likely other platforms to prominently display an image that is nowhere in their article. Plus did they ask Tommy Lee Jones permission to use his face to market their articles in social media? Presumably n
#
aaronpk lol
[kevinmarks] joined the channel
#
tantek__ aaronpk, are they doing that for other articles too? e.g. where you saw that list view
#
aaronpk looks like the more recent ones show the image on the blog post
#
aaronpk seems to have changed around march 2016
#
tantek__ nope I found one from 2018
#
tantek__ https://www.theregister.co.uk/2018/01/08/micron_and_intel_split_3d_nand_development/
#
aaronpk huh weird
#
aaronpk no idea what the pattern is then
#
tantek__ <meta property="og:image" content="https://regmedia.co.uk/2015/04/17/paltrow_teaser.jpg?x=1200&y=794"/>
snarfed joined the channel
#
tantek__ Is Gwyneth Paltrow a spokesperson for Micro and/or Intel now? Or a 3D chip thoughtleader? Perhaps we should check goop.com for NAND technology.
#
tantek__ Just need a few more to establish a pattern, misusing celebrity images to market their articles
#
tantek__ oh and all enabled by invisible metacrap, AKA dark data
#
snarfed double checks his IRC channel
#
snarfed 😂
#
tantek__ snarfed, started with an exploration of use of proprietary meta markup that's been debated past few days
#
snarfed yup, just a joke
#
tantek__ BTW I'm saving all these to archive.org just in case they change
#
gRegorLove Maybe Goop has an article about a cleanse for your computer. ;)
#
snarfed out of curiosity, you said they're doing this knowingly. how do you know? as opposed to just a bug, rotting invisible metadata, etc?
[cleverdevil] joined the channel
#
tantek__ snarfed good point! even better to express that as a possibility, and thus another downside of such "dark data tech"
#
[cleverdevil] A jade egg that fits in a USB port?
#
tantek__ [cleverdevil]: you can probably get jade-encrusted USB sticks
#
snarfed "never attributed to malice that which can be adequately explained by stupidity"
#
tantek__ snarfed, my first theory is that they're using the images in a meme-like fashion. the image of Tommy Lee Jones looks skeptical
#
tantek__ but the GP image I have no idea why
#
snarfed cache bug, fallback bug, other possible bugs
#
tantek__ in the first example we can posit some possible intent, the second example does illustrate your point, likely an error, rotting invisible metadata
#
snarfed anyway
[ryan339] joined the channel
#
[ryan339] i think one question is this actually malicious behavior? (ignoring the celebrity thing which is a red herring IMO). given that ogp/hidden meta tags are the only way to get these link previews on silos, is it actually wrong/bad/malicious to show an image in a link preview thats not in the article?
#
tantek__ deceptive / misleading, yeah
#
aaronpk I suppose it could be considered emergent behavior
#
aaronpk but that tag was definitely intended to show a preview of the article
#
tantek__ behavior that is certainly enabled by dark data
#
[ryan339] “link preview” is perhaps a misnomer here… its an “ad” to convince you to click on the link..
#
aaronpk it was created as a link preview initially
#
tantek__ [ryan339]: no those technologies were created to show previews of the content
#
[ryan339] true, i think it does go against the oriignal intention of the tag
#
tantek__ that's in their documentation
#
snarfed ooh this might be my first opportunity to invite people to go *from* #-dev *to* #indieweb ! :P
#
tantek__ thus it's an abuse, whether malicious or accidental
#
aaronpk snarfed++
#
Loqi snarfed has 71 karma in this channel (407 overall)
#
tantek__ their mis-use of celebrity personas is enough to raise red flags among a broader crowd
#
tantek__ and/or others' copyrighted imagery without attribution etc.
#
tantek__ here's another, pretty recent: https://www.theregister.co.uk/2018/06/26/kaspersky_transparency_tour/
#
tantek__ <meta property="og:image" content="https://regmedia.co.uk/2015/05/22/tough_mudder.jpg?x=1200&y=794"/>
#
tantek__ TOUGH MUDDER® is a trademark of Tough Mudder Incorporated. https://toughmudder.com/terms-of-use
#
gRegorLove OGP isn't the only way to get the link preview on Facebook, at least. Twitter, I'm not sure.
#
tantek__ that one also seems deliberate due to the title: "On Kaspersky’s 'transparency tour' the truth was clear as mud" - unlikely an accident they used a Tough Mudder image as a teaser for an article headline that mentions "mud"
#
[ryan339] maybe a better question: are there legitimate use cases for wanting to display a link differently in different contexts and thus needing to “abuse” ogp/card tags since thats the only option silos give us? (not saying these are legitimate use cases)
#
tantek__ [ryan339]: that's a misconception - it's not the only option - certainly not for FB
#
tantek.com edited /The-Open-Graph-protocol (+29) "/* Required Properties */ emphasize nothing is required" (view diff)
#
tantek.com edited /The-Open-Graph-protocol (+273) "make it clear you do not need og:title, url, or description for starters and why. list some reasons why you still might want to use them" (view diff)
#
tantek.com edited /The-Open-Graph-protocol (+127) "/* How to set image */ clarify you do not need og:image" (view diff)
#
tantek__ [ryan339]: here are the details on debunking the misconception that "og" is "the only option" (tag by tag) : https://indieweb.org/The-Open-Graph-protocol#How_To
[jgmac1106] joined the channel
#
tantek.com created /clicks (+25) "r" (view diff)
#
tantek.com created /dating (+33) "r" (view diff)
#
tantek.com created /friendships (+23) "r" (view diff)
#
gregorlove.com edited /advertising (+4) "/* IndieWeb Examples */ link The Deck" (view diff)
#
aaronparecki.com edited /weight (+133) "+indiewebcat" (view diff)
#
[jgmac1106] Register, bad headline.. Register should have done something much snarkier if they are throwing shade, good lede, gets sloppy where people usually stop reading
#
tantek.com edited /weight (+2) "fix dfn . link" (view diff)
#
loqi.me created /medical (+275) "prompted by tantek__ and dfn added by tantek__" (view diff)
[matpacker], snarfed, tantek and benwerd joined the channel
#
tantek__ Alright I need to distinguish in-person vs remote RSVPs on my IWC SF event for food ordering
#
tantek__ so that's my use-case, need some way to publish RSVPs differently accordingly
#
aaronpk they're separate on the wiki already, do you mean your indie event?
#
tantek.com edited /MediaWiki:Sidebar (+0) "808 (not just a State)" (view diff)
#
tantek__ aaronpk: yes, as in a use-case to drive enhancing /RSVP
#
tantek__ markup
#
tantek.com edited /rsvp (+1033) "add dougbeal example stub, brainstorm a few mf2 possibilities for remote RSVPs" (view diff)
#
tantek__ aaronpk, others interested, please review: https://indieweb.org/rsvp#Remote_Participation for possible ways to expand the "rsvp" property to handle remote participatnts
#
tantek__ !tell dougbeal can you expand https://indieweb.org/rsvp#Doug_Beal ? When was your first RSVP post? Etc. any other details you'd like to add
#
Loqi Ok, I'll tell them that when I see them next
#
aaronpk I thought there were already notes on this
[kim_landwehr] joined the channel
#
tantek__ aaronpk, there were minimal notes, I expanded a bit
#
tantek__ if you know of notes elsewhere (maybe lost in session notes that never got re-integrated into brainstorming?) definitely LMK!
#
tantek__ and next I'm going to wonder about dietary restrictions / preferences
#
tantek__ again, use-case is ordering food for an event
snarfed and [manton] joined the channel
#
[manton] [cleverdevil] I'm experimenting with adding a JSON Feed to complement the HTML archive page. It has the same content (truncated text for each post), but might be a little faster to parse. Unfortunately it's not smaller in terms of bytes. http://cleverangel.org/archive/index.json
#
dougbeal tantek__: sure
#
Loqi dougbeal: tantek__ left you a message 14 minutes ago: can you expand https://indieweb.org/rsvp#Doug_Beal ? When was your first RSVP post? Etc. any other details you'd like to add
[cleverdevil] joined the channel
#
[cleverdevil] Its worth a look.
#
[cleverdevil] I did some performance testing, and 90% of the processing time is using mf2py to parse the archive page.
#
[cleverdevil] If I had a JSON file and its equivalent in HTML, I could test in my environment to see if it would help!
#
[manton] You should be able to grab that link above and compare to the HTML version, then. Let me know if I can help.
#
[cleverdevil] Oh I totally missed the lionk
#
[manton] Yeah, sorry... When I said "I'm experimenting with", I really meant "It's live now". 🙂
#
[cleverdevil] 😉
#
tantek ^^^ lolol
#
tantek manton++
#
Loqi manton has 14 karma in this channel (76 overall)
#
tantek what is testing
#
Loqi Tests are a way to automatically check ones implementation for edge cases where it may break https://indieweb.org/Testing
#
tantek what is production
#
Loqi It looks like we don't have a page for "production" yet. Would you like to create it? (Or just say "production is ____", a sentence describing the term)
#
tantek production is /deployment
#
Loqi ok
#
loqi.me created /production (+23) "prompted by tantek and redirect added by tantek" (view diff)
#
tantek (close enough)
#
[cleverdevil] LOL, parsing the JSON is nearly twice as *slow*.
#
tantek WAT
#
tantek (!!!)
#
[manton] Ack! That is surprising.
#
[cleverdevil] I may be missing something.
#
[cleverdevil] Ah, yes I am missing something.
#
[cleverdevil] Is the JSON being generated on demand, or is it a static file?
#
tantek deployment << [[selfdogfood#testing_your_code_in_production|testing your code in production]]
#
Loqi ok, I added "[[selfdogfood#testing_your_code_in_production|testing your code in production]]" to the "See Also" section of /deployment https://indieweb.org/wiki/index.php?diff=50440&oldid=17135
#
[manton] Static file. Should be basically the same size and download speed.
#
aaronpk whoa
#
[cleverdevil] Okay, further testing.
#
aaronpk I mean that is a huge page
#
tantek.com edited /selfdogfood (+93) "/* testing your code in production */ thanks manton" (view diff)
#
[cleverdevil] Its muuuuch faster in JSON
#
[cleverdevil] There was a weird network thing going on.
#
[cleverdevil] If this is already live everywhere, I am going to make a change to see how it impacts performance.
#
[manton] [cleverdevil] Oh, you should wait... It's live but only for sites that have posted today.
#
[cleverdevil] Ah, gotcha.
#
[cleverdevil] I could put a conditional in there that checks for it.
#
[manton] That'd be cool. I could update it everywhere but probably better to let it update gradually the next time people post.
#
[cleverdevil] Indeed.
#
[cleverdevil] I'll do that.
#
[manton] [aaronpk] Yeah, for a big site this page is really big, HTML or JSON. Ultimately probably need a more efficient way to do these archive pages (breaking it out by year/month).
snarfed joined the channel
#
[cleverdevil] This is going to be so gross to handle it in two different ways at the same time 😛
#
[cleverdevil] But, worth it for testing, haha.
#
dougbeal.com edited /rsvp (+681) "/* Doug Beal */ first rsvp" (view diff)
#
dougbeal.com edited /rsvp (+58) "/* Doug Beal */ add description of first rsvp" (view diff)
#
[manton] [cleverdevil] If things go well, I can probably update all the existing sites over the course of a few days, and you could remove the extra code later.
#
[cleverdevil] Yeah, that's what I was thinking.
#
[cleverdevil] Its no worries.
#
aaronpk [cleverdevil]: whatever happened with that mf2 json schema validator thing?
#
[cleverdevil] I have it up on github.
#
[cleverdevil] https://github.com/cleverdevil/microformats2\
#
Loqi [cleverdevil] microformats2: Validate Microformats2 JSON with JSON Schema, and perform Post Type Discovery
#
aaronpk that's a confusing repo name :P
#
[cleverdevil] It also does post type discovery and JF2 conversion
#
[cleverdevil] Yeah likely 🙂
#
aaronpk what does the validate() function return?
#
[cleverdevil] I believe it returns nothing
#
[cleverdevil] It raises an exception in the event of a validation failure
#
[cleverdevil] Likely should return True in the event of no errors./
#
aaronpk ah, is there any explanation in the exception?
#
[cleverdevil] Yes
#
aaronpk what would it take for you to throw that up on a little web service that takes a URL and displays the result?
#
[cleverdevil] I could make it available as a web service, if it would be useful.
#
[cleverdevil] Haha
#
[cleverdevil] Not much effort. I'll see if I can find some time!
#
[cleverdevil] In the meantime, [manton] I have deployed that change and now things are inexplicably slower for some reason.
#
[cleverdevil] My debug logs indicate it should be *faster*.
#
aaronpk also could you make it only accept the full parsed result? e.g. not accept just a single mf2 item?
#
[cleverdevil] I need to look into more detail.
#
aaronpk to make it require {"items":[ ... ]}
#
[cleverdevil] [aaronpk] it accepts either, I believe.
benwerd joined the channel
#
dougbeal.com edited /invitation (+365) "/* IndieWeb Examples */" (view diff)
snarfed joined the channel