#dev 2019-02-05

2019-02-05 UTC
tantek and eduardm joined the channel
#
[tantek]
I thought I saw something from jacky about posting an RSVP - but didn't find one on this week's HWC
#
Loqi
[Jacky Alcine] Excited to finally come to a website club meetup this week! Might be coming in with more questions and ideas and demo a bit of what I have.
[kevinmarks] joined the channel
#
[tantek]
oh no it's inside the h-cite so it's not showing up as part of the h-entry!
#
[tantek]
jacky ^^^
#
[tantek]
your <div data-rsvp="yes" class="dib mw4 tracked ttu br--right ph3 truncate f6 v-mid tc p-rsvp">yes</div> is inside of <header class="flex flex-row flex-wrap items-center pa2 h-cite u-in-reply-to"> which is not true, because your RSVP is not part of the thing you're citing, and also results in the rsvp not being found - which is expected as a property of the h-entry at the top
[cleverdevil] joined the channel
#
[cleverdevil]
Also, you can now search by name and username and nickname
#
[cleverdevil]
Also, it now works for non-domain identities like Zegnat's
#
[cleverdevil]
Looks like nicknames aren't persisting for some reason, though... not sure why
#
[tantek]
you may want to add some more URL parsing smarts to Twitter handle discover, as there are some top level segments that are not usernames, e.g. see https://pseudonym.cleverdevil.io/identity?url=http://tantek.com
#
[cleverdevil]
Nice catch!
jackjamieson joined the channel
#
[cleverdevil]
Nicknames are also fixed now.
#
gRegorLove
cleverdevil++
#
Loqi
cleverdevil has 43 karma in this channel over the last year (92 in all channels)
jjuran joined the channel
#
jacky
!tell tantek I'll fix that just now
#
Loqi
Ok, I'll tell them that when I see them next
jgmac1106 and [tantek] joined the channel
#
Loqi
[tantek]: jacky left you a message 4 minutes ago: I'll fix that just now
#
jacky
I think I broke it? I tried https://jacky.wtf and it 502'd
#
Loqi
jackyalcine
#
jgmac1106
jacky I was looking at your h-card what is n-nickname?
jjuran joined the channel
#
jacky
lol that's a mistake
#
jacky
can't wait to move away from jekyll for that
jgmac1106, jjuran, [tantek], [eddie], eddiehinkle and [cleverdevil] joined the channel
#
[cleverdevil]
Was afk for a bit. Glad it’s working mostly! I’ll check into what’s going on with your site in a bit jacky.
#
[cleverdevil]
Actually it looks like it’s working fine now 😀
#
jacky
<magic.gif>
#
[cleverdevil]
I think it’s because I am running it in turbo not production mode with a single threaded web server behind Apache proxy 😀
#
[cleverdevil]
Once I get it locked in a little more, I’ll make it not dumb.
snarfed joined the channel
#
Loqi
[cleverdevil] pseudonym: Web service for IndieWeb identity discovery, "nickname" search, and replacement.
#
jacky
_niceee_
#
jacky
[cleverdevil]++ x3
#
Loqi
[cleverdevil] has 44 karma in this channel over the last year (93 in all channels)
#
[cleverdevil]
Loqi, give me my two more!
gRegorLove_, [Rose], cweiske, eduardm, gRegorLove__ and [tantek] joined the channel
#
petermolnar
I caved in. I've had enough of trying to h4xx0r rdfa into HTML and added a json-ld into my site :(
tglobe and tglobe_ joined the channel
#
Zegnat
I thought you had the RDFa down? Wasn’t it just a copy of the microdata you were already shipping?
[tantek] joined the channel
#
petermolnar
it had issues
#
petermolnar
microdata was straighforward, especially with the itemref for author and such
#
petermolnar
rdfa is madness
#
petermolnar
eg. I couldn't get google to parse emails properly
#
petermolnar
so the short summary is that microdata is sort of copy of mf2, with different property names
#
petermolnar
rdfa is a nope
#
Zegnat
Google bug or RDFa limitation on that “madness”?
#
petermolnar
maybe both
#
petermolnar
as said, I gave up
#
[tantek]
sorry you had to experience it firsthand petermolnar 😞
#
petermolnar
nah, it's fine
#
petermolnar
I did this to myself :D
#
petermolnar
on the other hand, json-ld is relatively straightforward, but it annoys me a I need to add my author block a good couple of nested times
#
petermolnar
eg. if I have a Photograph, which has an ImageObject as, both needs my author as author, creator, copyrightholder
#
petermolnar
and I can't simply say author == creator == copyrightholder
#
Zegnat
Huh? I thought JSON-LD came with object references
#
petermolnar
how?! tell me! :D
#
Zegnat
Hang on, trying to find out
#
Zegnat
Though I really thought I had seen examples of this within a single JSON-LD document
#
Zegnat
Because you don’t really want to work with external stuff
#
petermolnar
unless it's something like self@author, it's not really what I'm after
#
Zegnat
If you embed an @type=Person on the first author thing, and pass it a document relative @id (starting with # in this case, not path or domain relative), other places where you want to refer the same person could just be {"@id":"#whatever"}, I guess.
#
Zegnat
Though not sure if that is actually correctly derefed by any parser
#
Zegnat
Hmm, you may be able to use blank nodes and refer those, if you like mixing RDF in your LD :P
#
Zegnat
Not joking, clicking the term "blank node" in the JSON-LD spec literally links you away to the “RDF 1.1 Concepts and Abstract Syntax”
#
petermolnar
is nearly finished piling the logs to burn anything related to computing around him
#
petermolnar
Recursive Descent Forever
#
petermolnar
[kevinmarks] was spot on
jeremych_ joined the channel
#
[tantek]
petermolnar did you keep track of how many hours you spent on this, and what meaningful/useful results you got (if any?)
#
petermolnar
no, I don't see a point in it, it's for my own (g|p)ain
#
[tantek]
point is to record your experience as a warning to others
#
petermolnar
I can summarize the warning: microdata or json-ld
#
petermolnar
ignore the rest
eduardm_ joined the channel
#
[tantek]
even with those, how many hours vs how much incremental additional useful functionality did you get?
#
[tantek]
would be useful to get a modern debunking of RDFa documented by someone new
[kevinmarks] joined the channel
#
[kevinmarks]
We spent a lot of time trying to get the 'include this from elsewhere in the document' thing in mf1 and decided it was too complicated
#
[kevinmarks]
The pattern is relatively clean, but still tricky to get right.
#
Zegnat
[kevinmarks]: mostly a voice in the back of my head shouting that it is something JSON-LD can do, but then I can’t figure out where/how
[jgmac1106] joined the channel
#
[kevinmarks]
With mf2 it could be clearer because we have explict embedding of a h- within a named property, whereas all the nesting in mf1 was vocabulary specific knowledge
#
[kevinmarks]
So if we did define something like include pattern for mf2 we could use a url. However you are then implicitly requiring async parsing (the include is a bit like a promise).
#
Zegnat
In the current case, as I understand it from petermolnar, the included resource would be in the same document, which does make parsing a little easier as it removes the need for fetches
#
Zegnat
petermolnar: here you go! Blank nodes can be used to have nested objects include/reference another one: https://www.w3.org/TR/json-ld/#identifying-blank-nodes (see contrived example)
#
[kevinmarks]
Yes, but it's still a dangling reference so you can't just use recursion.
#
petermolnar
this is what I actually want
#
petermolnar
I'm confused
#
Zegnat
That is just for renaming properties
#
Zegnat
e.g. if you want to use "author" everywhere you can alias "copyrightholder" as "author"
#
petermolnar
right; so it's your find
#
petermolnar
the one I actually wanted
#
Zegnat
Probably. I *think* so. But take it all with a healthy bucket of salt. This isn’t really my expertise
wagle and globbot joined the channel
#
Zegnat
apologises to petermolnar for being a JSON-LD-enabler :P
#
[jgmac1106]
Was playing with it as well last night ( I think could have just been useless javascript as well) https://quickthoughts.jgregorymcverry.com/2019/02/05/anomalily-anildash-i-tried-to-make-you
#
Loqi
[Greg McVerry] @anomalily @anildash I tried to make you a quick business card https://bird-society.glitch.me/ but kept running into data that needed to go into javascript rather than HTML where it belongs so I gave up.
#
@jgmac1106
I turned native comments off my main site and rely just on webmentions. These are archived sites I do not use any more. Don't feel as bad. (https://quickthoughts.jgregorymcverry.com/s/WSqtv)
(twitter.com/_/status/1092748454814773248)
#
Zegnat
I don’t understand, [jgmac1106], what were you trying to use JS for on a static user card?
#
[jgmac1106]
No I was making a fast business card as a fast joke and the template I randomly opened needlessly put useful data in javascript
#
[jgmac1106]
I mean I guess having all your bio deets in a javascript file you can call whenever would save some time... But meh... Gimme readable HTML
#
petermolnar
while I dislike it, json-ld is quite readable
#
petermolnar
I'd like to help and see the problem you saw
#
Zegnat
“json-ld is quite readable” and parsable, which is probably the reason Google likes it
#
[jgmac1106]
Didn't see a problem at all, just wanted to write the HTML it was a great excercise though in understanding why nickname cache gets stored injson-ld
#
[jgmac1106]
My only problem was speed and my own knowledge.. And just a chance to promote an HTML first world
#
[jgmac1106]
And for other people. I want to see words and not scripts when I hit source code.. I can't stand websites I don't know how to read yet bc of wizardry
#
Zegnat
I don’t think anyone is storing nickname cache in JSON-LD?
#
[jgmac1106]
No sorry meant JSON.. Auto completed
#
[jgmac1106]
Means I have typed JSON-LD too much lately
#
petermolnar
json vs json-ld is a matter of @context
#
Zegnat
Also a couple of other rules. Not all JSON documents are valid JSON-LD documents (depends on top-level object and stuff)
#
Zegnat
But JSON makes a lot of sense over HTML if it isn’t human-facing data. It sure is a lot easier to parse.
#
petermolnar
I've seen JSON more human readable, than HTML, reading is not the problem of JSON, writing is
#
Zegnat
I am thinking if you get HTML to read, you will read it using a browser
jgmac1106 and iasai joined the channel
#
jgmac1106
what does facet mean in this context, just a form field? https://jonudell.info/h/facet/
bradenslen and jgmac1106 joined the channel
#
sknebel
re your edit: oh, they actually switched to the W3C annotation format? interesting
#
jgmac1106
I think they were the only one’s pushing for the W3C annotation format and I can’t find an example of anyone else using it
#
jgmac1106
I looked hard this morning but did not find anyone
#
sknebel
I seem to remember there'd been a few other (non-commercial) projects working with it, and hypothesis had been making big blogposts about it but hadn't at that point actually done much with it - but it's been a while since I looked at that
[kevinmarks], cjwillcock, iasai, iasai_, jgmac1106, eddiehinkle, swentel, [davidmead], [svandragt], test, KartikPrabhu, barpthewire, [tantek], snarfed and [jgmac1106] joined the channel
#
@digitals_cat
Implementing Web Mentions. How to claim back mentions from social media https://peeke.nl/implementing-webmentions
(twitter.com/_/status/1092822704384806913)
eli_oat and [schmarty] joined the channel
#
snarfed
!tell cleverdevil congrats on launching pseudonym! just fyi, https://pseudonym.cleverdevil.io/identity?url=https://snarfed.org doesn't find my h-card. i skimmed Identity.fetch() but haven't yet seen an obvious bug
#
Loqi
Ok, I'll tell them that when I see them next
#
[schmarty]
snarfed: it seems to work when i load it. maybe an intermittent issue?
#
snarfed
[schmarty]: it finds the rel-mes, not the h-card
jgmac1106, tantek_, [chrisaldrich], eddiehinkle, [gerwitz] and [cleverdevil] joined the channel
#
Loqi
[cleverdevil]: snarfed left you a message 1 hour, 11 minutes ago: congrats on launching pseudonym! just fyi, https://pseudonym.cleverdevil.io/identity?url=https://snarfed.org doesn't find my h-card. i skimmed Identity.fetch() but haven't yet seen an obvious bug
#
[cleverdevil]
So, yes, mf2py doesn't find your h-card.
#
[cleverdevil]
Well, it does, but its embedded inside of your h-feed, but then doesn't find it when filtering by type.
#
[cleverdevil]
I really just need a better algorithm for discovering the representative h-card for a particular site.
bradenslen, jgmac1106, iasai, [jgmac1106] and [davidmead] joined the channel
#
[cleverdevil]
Someone who better understands representative h-card should add this capability to mf2py 😛
#
sknebel
[cleverdevil]: mf2utils should do that
#
sknebel
sorry, mf2util
#
[cleverdevil]
Oh? I wasn't aware of mf2util
#
[cleverdevil]
I'll go find it 🙂
jgmac1106 joined the channel
#
jgmac1106
What is mf2util?
#
Loqi
mf2util is a Microformats 2 utility for Python, commonly used to extract common features in comments and reply-contexts https://indieweb.org/mf2util
#
jgmac1106
found it
#
[cleverdevil]
Unfortunately, it looks like it doesn't work particularly well either.
#
[cleverdevil]
My poor algorithm is more reliable thus far.
eli_oat joined the channel
#
[cleverdevil]
Ah, undocumented function works pretty well.
#
[cleverdevil]
And in the gaps I can fill them in.
#
[cleverdevil]
Deployed an improved version, should find most everyone's h-card now 🙂
[pfefferle] and [eddie] joined the channel
#
@jgmac1106
↩️ Also @WickedEdTech if you want me to show you how to replace Google+ comments (before 4/2) with webmentions I can some time. Webmentions are wicked cool! (https://quickthoughts.jgregorymcverry.com/s/1TINgk)
(twitter.com/_/status/1092864348291194885)
#
sknebel
[cleverdevil]: oh, can you file an issue for missing docs?
swentel joined the channel
[kevinmarks] joined the channel
#
[cleverdevil]
Actually, its documented, I just missed it 😉
[Rose] and snarfed joined the channel
#
snarfed
mf2util++ , i use it in granary and bridgy too. yeah the docs are a bit spotty but it's good software
#
Loqi
mf2util has 3 karma over the last year
snarfed joined the channel
#
KartikPrabhu
might be good to reach out to kylewm to have mf2util either in the microformats or indieweb repo
#
snarfed
gof ror it!
#
[cleverdevil]
It also has post type discovery, which I wasn't aware of when I implemented it myself in my own microformats2 package.
#
[cleverdevil]
May be smart to merge all of these things into a single utility at some point...
#
KartikPrabhu
not a bad thing to have multiple implementations
#
snarfed
yeah mf2util is the pretty clear candidate; it has most (all?) things already
#
snarfed
eh. diversity vs reuse. yeah there's a tension.
#
snarfed
we have enough different languages w/their own implementations that reusing within languages is reasonable
#
snarfed
but yeah moving mf2util into the mf or iw org is a good idea, and would signal more clearly to reuse it
eli_oat1, [schmarty], KartikPrabhu, snarfed, tantek, tantek_, [eddie], [jgmac1106], [kevinmarks], iasai, [dougbeal] and dougbeal|mb1 joined the channel
#
jacky
from -chat, I want to pull in the music I listen to on my site
#
jacky
but this is my yearly graph https://i.imgur.com/2XGEMzr.png
#
jacky
I don't even know how I did so much in 2018
#
jacky
(obvs my place is never quiet lol)
[kevinmarks] joined the channel
#
aaronpk
Yepppp
#
aaronpk
this is why I haven’t done that yet either
#
aaronpk
pretty sure I want to treat it differently from my regular posts because of the volume