#dev 2020-04-06

2020-04-06 UTC
mastermidn13, [tantek], geoffo, [LewisCowles] and [KevinMarks] joined the channel
#
[KevinMarks]
now I get to work on the visual display properly
superkuh, [Jeff_Hawkins] and KartikPrabhu joined the channel
#
crab
so it's not entirely trivial to get schema.org RDFa markup to coexist with h-entry/h-card markup.
swentel joined the channel
#
[tantek]
what's the consuming use-case for RDFa markup?
KartikPrabhu joined the channel
#
swentel
I guess mostly rich snippets in search results?
#
Loqi
swentel: jacky left you a message 2 days, 8 hours ago: it works nicely!
#
[tantek]
swentel, for rich snippets in search results, all you need is a top level hentry / hcard, classic mf1 which works great in-place with mf2: http://microformats.org/wiki/faq#When_should_I_use_microformats2_or_microformats1 (no need for RDFa for that)
#
Loqi
Microformats FAQ
#
swentel
[tantek], I know, but, I guess that's it's 'sold' for, at least by marketing :/
#
crab
[tantek] for me, mostly just curiosity
#
crab
i remember a time when RDF was still the great hope for semantic markup :-)
#
crab
having added h-entry/h-card/hCard markup to my site already, i just wondered if it would be easy to also markup as BlogPosting/Person with RDFa, which google claims to pay some attention to
#
crab
and it's not too difficult, though there are some oddities if you want to use rel=author. but google's structured data testing tool says the article MUST have an image, which most of mine don't, so i probably won't bother.
loicm and [LewisCowles] joined the channel
#
[LewisCowles]
well done KevinMarks
#
[LewisCowles]
Cheuk, can you describe the break in more detail?
#
[tantek]
very odd about requiring an article to have an image, that doesn't make sense as most blog posts do not have images
sscarfe joined the channel
#
[LewisCowles]
is it the standard or Google with that requirement?
#
[LewisCowles]
I feel like it's the latter.
#
[tantek]
anything "rich snippets" is a Google proprietary request
#
[tantek]
sometimes they re-use some standards in those requests
#
[LewisCowles]
sad but familiar
#
crab
[LewisCowles] it's a google requirement (image and publisher)
#
crab
oh, sorry, tantek already said so.
#
crab
and the annoying thing about making hCard/h-card/Person markup coexist is that google's structured data testing tool doesn't like "rel=me".
#
crab
so, blah.
[ColinMorris] and [KevinMarks] joined the channel
#
[KevinMarks]
pours one out for the social graph api
#
Loqi
Google has -3 karma in this channel over the last year (-9 in all channels)
#
[LewisCowles]
just found out lit-html doesn't live up to the cool "that {x} was lit"
#
[tantek]
I hope that's u-in-reply-to
#
Ryuno-Ki
Regarding RDFa, I could be wrong, but wasn't SoLiD (by Sir Tim Berners-Lee) using it in some way as well?
#
Ryuno-Ki
The Yandex Webmaster tool were more helpful for analysing mf2 / JSON-LD and that like. But they require login now :-(
#
Ryuno-Ki
Fun think is that even the schema.org examples don't validate …
#
Ryuno-Ki
At least solid is suggesting use of rdflib: https://solid.inrupt.com/docs/manipulating-ld-with-rdflib
gxt, sscarfe, inkjama, [jgmac1106] and bmat joined the channel
#
[jgmac1106]
SOLiD does use RDFa, but I don't know any SOLID users yet
#
[jgmac1106]
I think there are a few community members exploring it, youcan check the wiki page
hs0ucy joined the channel
#
[jgmac1106]
almost all the major news organizations and academic journals doing some version of JSON-LD and schema.org vocabularies..for this hot minute
KartikPrabhu joined the channel
#
crab
schema.org vocabularies seem pretty all right
sscarfe joined the channel
#
crab
but JSON-LD is annoying. i can imagine why google likes it, but i much prefer the microformats approach
#
Ryuno-Ki
Same here, crab
#
crab
anyway, i ended up adding schema.org/BlogPosting markup in RDFa format to my site, because it was a five-line diff in the end. but i ignored all the extra stuff (image, publisher) that google's validator asks for.
#
crab
any more microformats and it'll become a miniformat.
#
Ryuno-Ki
I think, the requirements comes from a marketing perspective (posts with images tend to attract more eyes).
#
crab
i think they ask for the image because they want to feature it in this "carousel" of results as an image with text underneath
#
Ryuno-Ki
Could be. But afaik Facebook/Twitter/Slack etc. are using it for the thumbnail preview (if you haven't add their markup already)
[grantcodes] and sscarfe joined the channel
#
[grantcodes]
[jgmac1106] if you want to be super experimental with your scaling font size you could try `clamp()` in css which lets you set a min and max value
#
[jgmac1106]
thx Grant, I started to read about min/max and vm fontsizes, ended up doing wiki stuff instead. I do need to do something for my small screen font sizes
KartikPrabhu joined the channel
#
petermolnar
schema.org vocab is nowhere near all right
#
petermolnar
try making a resume in it
#
petermolnar
or recurring events
#
jgmac1106_
not about being right, about academics making a career proposing vocabularies that probably won’t be used much,
[LewisCowles] joined the channel
#
[LewisCowles]
↩️ hmm, sadly not a lot of insight in that message is there. Have you emailed them?
loicm, sscarfe and [grantcodes] joined the channel
#
jgmac1106_
does anyone see my markup mistake? Telegraph is telling me no h-entry is present on this post: https://jgregorymcverry.com/lofilearningmozillagrant
#
Ryuno-Ki
However, OpenLink Structured Data Sniffer can't show me details either
KartikPrabhu and loicm joined the channel
#
aaronpk
jgmac1106_: possibly because it's missing a url property
#
aaronpk
tho I don't remember why Telegraph expects that
#
jgmac1106_
duh…I probably never updated my article template….one typo haunts me over and over
#
jgmac1106_
I thought maybe it was the extra h-card I had…
#
aaronpk
Ah it's because you have more than one h-* object on the page and there is no url property to confirm which is the primary one for the page
#
sknebel
which UI of telegraph does expect it? I put it in the "send webmentions for this link" UI after login and it showed me the links on the page?
#
jgmac1106_
sknebel after you send it fails
#
jgmac1106_
removing the h-card on the photographer
#
sknebel
telegraph, or the webmention endpoint you tried to send to?
#
jgmac1106_
I did think about wrapping the entire photo as it’s own h-entry since it has a different license and author,
#
aaronpk
It's your h-card in the bottom. The fix is to add the url property in the h-entry
#
jgmac1106_
ahh okay, fixed that too
#
aaronpk
tho this case has come up a few times now and I wonder if I should accept it.
#
jgmac1106_
thank you both…hopefully now all my templates are fixed
#
crab
i always thought the url in the h-entry was required
#
aaronpk
specifically a single h-entry and a single h-card is unlikely to be a feed, which is how it comes out of XRay right now
#
sknebel
if you still have a tab open, curios what the exact error was
#
aaronpk
sknebel: i think telegraph does another verification before it sends, and apparently the two ways it verifies don't match
#
sknebel
then why does it not show the error when I tried jgmacs url?
#
aaronpk
Cause he fixed the markup I think?
#
sknebel
I tried before
#
aaronpk
And you were able to have Telegraph send the webmentions?
#
sknebel
and then the indienews endpoint explained to me what the problem with the markup was
#
aaronpk
That makes more sense, I didn't think telegraph cared much what a page is, just that the link exists
#
sknebel
aaronpk: also re "should accept that" - if you come up with a proper definition of what an acceptable permalink is then that everyone who is not using xray can follow
#
sknebel
to replace the "has u-url to current page" definition
#
sknebel
but that's why I'm curious what exact error jgmac1106 saw
#
aaronpk
I was going off of what XRay was returning from that url
#
jgmac1106_
let me check skenbel, it was on the indienews link only not indeweb.xyz or my endpoints
#
aaronpk
What is this exactly? It's not post type discovery
#
Loqi
It looks like we don't have a page for "this exactly" yet. Would you like to create it? (Or just say "this exactly is ____", a sentence describing the term)
#
aaronpk
Hey look at that the error message explains exactly what I said
#
aaronpk
"If you have multiple top-level h-* objects, ensure that one of them has a u-url property set to the URL of the page"
#
jgmac1106_
yeah once I expanded my broswer window I could read it
#
jgmac1106_
can you had an overflow or break the text..it is my mistake but I could not read the error, eveything is there to fix it
#
aaronpk
Yeah what's the right way to format JSON with word wrapping? Simple wrapping would look weird IMO
#
jgmac1106_
I probably didn’t look past the cur off, once I read no h-entry found
#
jgmac1106_
can you you just had an overflow scroll? need to ask someone with more skill than me
#
jgmac1106_
the table is correct, not sure why it didn’t scale
#
sknebel
and this is why describing exactly what you see when asking for support is important
#
jgmac1106_
good point, thanks, will remember. thx for the help, collecting webmentions as endorsements is fun
#
jgmac1106_
on another ntoe redoing elmcitywebmakers.com as we are moving it to a statewide online program, I want to pay homage to IWC colors but these burn eyeballs: http://elmcitywebmakers.com/
#
jgmac1106_
normally I just type in a hex number and randomlu choose a color pallette on a hue
#
jgmac1106_
but if anyone has any color ideas this is right in the spectrum where I don’t make sense of anything and have to color by number
[jgmac1106] joined the channel
#
sknebel
petermolnar: bad news: that cloud provider is probably cloudflare, or the free quota of some "serverless" product
#
petermolnar
cloudflare :/, serverless? serverless! (backstory: a colleague of mine got a bit frustrated with the term, he ended up responding "serverless!" as a reply to nearly anything for a day)
[KevinMarks] joined the channel
#
[KevinMarks]
found a bug in13-year-old utility routine, sent a PR
#
[KevinMarks]
[petermolnar_] can you say JAMstack
deathrow1 and [LewisCowles] joined the channel
#
[LewisCowles]
petermolnar: I love your colleague. The inappropriate and appropriation of perfectly legible English, into pure garbage; almost seems a goal by some movements.
#
swentel
hmm, aaronpk could you check if I still have an account on aperture?
#
swentel
I basically forgot the number :)
#
swentel
s/number/id
#
GWG
swentel: I tried your desktop app
#
GWG
Enjoyable.
#
GWG
Although it revealed some things I need to fix
#
[grantcodes]
jamstack isn't serverless
#
swentel
GWG, oh, there might be some bugs in the app too of course ..
#
swentel
GWG, what's your microsub server? aperture, or wp?
#
GWG
swentel: Never tried Aperture honestly
#
GWG
swentel: Only Yarns, which is the WP one
#
swentel
GWG, ok
#
Ryuno-Ki
jgmac1106_: regarding color palette: perhaps https://brandcolors.net/ could give you an idea
#
GWG
jackjamieson has just been busy, I think.
#
[grantcodes]
I think the netlify name of "cloud functions" is less misleading 😛
#
GWG
I keep saying I should send some PRs
#
Ryuno-Ki
Well, Function-as-a-Service (FaaS) and Serverless aren't the same either from what I understand.
#
swentel
GWG, I have this issue with the desktop app: https://github.com/swentel/indigenous-desktop/issues/45 uses wp for micropub, but aperture for microsub. And he's getting 403 on the microsub call .. trying to think what might be wrong there.
#
Loqi
[stephenpieper] #45 Microsub Channels
#
GWG
I am not. It may be the same issue he would have with any IndieAuth service
#
GWG
I'd direct him to the IndieAuth plugin for supportg
#
swentel
might have to dive in aperture code, need to check how authentication goes
#
GWG
If he's using the Aperture plugin for WordPress, it needs the IndieAuth plugin
#
swentel
GWG, even though token/indieauth discovery are set to wp?
#
GWG
swentel: The IndieAuth plugin handles the endpoint, so if he's getting forbidden, it should be an issue there
#
swentel
hmm, I see aperture uses a cache for tokens
#
swentel
maybe he entered another token
#
swentel
which aperture doesn't know about
#
swentel
or, the token doesn't have the right scopes if a new is generated in wp (if that's possible?)
#
swentel
mm, doesn't look like it
#
jacky
now I need a micropub client capable of sending mf2 for a 'page' (marked up as `h-x-page` for now, only accepting `content` and `name` for now)
#
swentel
ok, installing aperture plugin :)
[Michael_Beckwi] joined the channel
#
[grantcodes]
jacky: Won't any client with edit functionality work?
#
jacky
yeah but I'm talking about sending it for creatioin
#
jacky
*creation
#
jacky
https://micropub.net/draft/#create marks it as `h=entry` by default and most (if not all) clients focus on that
[snarfed] and [Jeff_Hawkins] joined the channel
#
[grantcodes]
Ah, why don't you use entry? I'm sure people have already done it that way
#
jacky
I haven't seen much on /page or /custom_page in terms of implementation
#
jacky
and I'm experimenting with it under that type
#
jacky
there's things I'd want a entry to have that I don't want pages to have (syndication targets, categories, etc)
[jeremycherfas], gRegorLove, bmat and inkjama joined the channel
#
[LewisCowles]
well for some reason that led me down a path... https://indieweb.org/database-antipattern was found...
inkjama joined the channel
#
[LewisCowles]
do people stand-by that?
#
[KevinMarks]
there are things databases are good for; storign pages may not be that.
#
[jeremycherfas]
!tell aaronpk Is there a way to see contents of a POST created by Quill?
#
Loqi
Ok, I'll tell them that when I see them next
#
[LewisCowles]
I'd strongly ask you to consider if file-systems are a great solution for that either. It's very XKCD
#
aaronpk
[jeremycherfas]: not built in
#
aaronpk
i think zegnat made something to help debug though
#
aaronpk
what is sink?
#
Loqi
Sink is an experimental site by Martijn van der Ven that allows anyone with an IndieAuth enabled URL to post to it using any Micropub client https://indieweb.org/Sink
#
aaronpk
[LewisCowles]: databases are good at indexing, but i prefer to store my posts in files since they're a lot easier to back up and transfer and maintain
#
[jeremycherfas]
Or I guess I could look in the console?
#
aaronpk
i dont' think it debugs it to the console
#
aaronpk
and the server makes the micropub request, so it's not necessarily what you see in the network tab either
inkjama joined the channel
#
[jeremycherfas]
OK. But it is a start. I’m pursuing the idea floated by ReneM[m] of using IFTTT or Zapier to PESOS photos from IG to WithKnown.
#
[LewisCowles]
[aaronpk]: it's not just indexing. Most take steps nobody here is taking to work around OS limits, but also, most are more standard and cross-platform than file-systems.
#
[LewisCowles]
do you use HFS, or HFS+, is it case-sensitive or not?
#
[LewisCowles]
Also at the point you've got a consistent file-system, even if you revise the structure going forward, it's just two schemas. You're less portable and standard than most db's
#
aaronpk
it's also one of the reason static site generaters are so popular
#
aaronpk
turns out it's easier to manage a pile of files than keep up with updates to database software 🤷
#
[LewisCowles]
oh man. so none of this is outside of opinion
#
aaronpk
everything is opinion
#
aaronpk
you're just sharing your own opinion too
#
[LewisCowles]
no I'm not. I've shared facts about file systems being inconsistent due to config
#
aaronpk
if you have specific examples one way or the other feel free to add them to that wiki page, which is a giant list of examples on both sides
#
[LewisCowles]
I've challenged the notion of no database being a file-system
#
[LewisCowles]
because the kind of file-system you'd need to keep would be a poor persons database
#
aaronpk
"need" is relative
#
aaronpk
plenty of people use filesystems to store data successfully
[tantek] joined the channel
#
[LewisCowles]
the argument you're making is why to put everything micro-formats into p tags and divs with no HTML classes#
#
aaronpk
no...who said that
#
[LewisCowles]
that's the logical conclusion of the points you just made
#
aaronpk
not at all, that's..i forget the name of that argumentative technique
#
KartikPrabhu
eh! no I use mf2 to store HTML files which are completely "semantic"
#
aaronpk
but like, if you're going to make that arugment then i'm done here
#
aaronpk
cause that's not a productive use of anyone's time
#
[LewisCowles]
exposing logical inconsistency. I was trying to get you to notice the behaviour, but I respect if you're done
TGiske joined the channel
#
[LewisCowles]
FWIW I think the post on the wiki belongs in the fire. It's needlessly devisive, uses examples of poorly used technology to argue against that technology, for something which is an imitation of that technology. It's a different thing saying you don't need a database all the time, to exploring when you might need one, and gotchas to avoid.
#
[LewisCowles]
this is search'es fault. I was trying to find /custom_page
geoffo, dietricha_, Kemwer_, [Jeff_Hawkins], TroyDiscord[m], talbDiscord[m], hubaDiscord[m], HuurooDiscord[m], thestevewayDisco and TGiske joined the channel
#
[tantek]
Ryuno-Ki, SOLID uses JSON-LD, not RDFa. And AFAIK no one (here or anywhere) actually uses SOLID like on the web to publish etc. Not even the team, they all directly use Twitter themselves. It's not an end user product, more like a developer exploration if you have a Linked Data hobby.
swentel joined the channel
#
[tantek]
lol [LewisCowles] discovered the database anti-pattern 🙂
#
Zegnat
Hmm, I have seen very few people with fully public and interesting linked data projects online. I really only know of Ruben’s
geoffo joined the channel
#
[tantek]
Zegnat, yeah it's pretty much academic. A few government things, which I have a feeling will no longer get the resources needed to keep them up to date etc.
#
Zegnat
Ruben’s data store: https://data.verborgh.org/ruben (thousands of LD triplets). Oh, and I do not remember having seen this, but there is a query tool aswell: http://query.verborgh.org/
#
Zegnat
Once you see it being applied that way with the querying, or just randomly digging through his triplets, I actually find it to be really fascinating
#
[tantek]
[LewisCowles] if you like databases, wait til you discover triple-stores 😉
#
[LewisCowles]
a type of database?
#
[LewisCowles]
I know what they are btw
#
[LewisCowles]
All a database is, is a way to organise data. Informal, like opt-in microformats, HTML, are still databases
#
[LewisCowles]
they just have fewer consistency parameters and features. you could argue the simplicity & permissive approach is desirable. That is not what calling a whole field of computing an anti-pattern is, regardless if you keep that to blog-posts, which have tags, categories, and other features that are difficult if not rolling your own database
#
[tantek]
the primary criticism, open maintainable standard (interoperable across multiplementations) file format, vs. not, still stands, and invalidates the "HTML is a database" assertion
#
[tantek]
the criticism is directed at database software, which is typically what people mean/imply when they say "database" rather than the CS definition of "database"
#
[LewisCowles]
FYI, I've been through the anti-databases phase of youthful rebellion. It didn't take much to come out the other side.
#
[tantek]
Nah, I did it the other way around, I've been through the youthful naivete of database files lasting, and came out the other side with sustainable open standard data formats instead
#
[LewisCowles]
Even the examples against MySQL were silly. We used it badly, or they were attempting consistent collation (which files don't) is more software iteration cost than anything
#
[jgmac1106]
plurality++
#
Loqi
plurality has 2 karma in this channel over the last year (4 in all channels)
#
[tantek]
Nope, even experts use MySQL badly and fail. That's the problem.
#
[jgmac1106]
you get a good mix of both here
#
[LewisCowles]
sustainable file formats. How do you access / enable large amounts of data?
#
[tantek]
When people fail with technology more often than not, it's not the people's failure, it's the technology's failure. Usability 101
#
[LewisCowles]
you'll hit file open limits if you don't make it into a pseudo database at some point
#
[tantek]
LewisCowles I have over 10 years of blog posts in static HTML files that work great!
#
[LewisCowles]
the second you start adding architecture tiers to mitigate that, you're building a database
#
[tantek]
Maybe you have different "large amounts of data" on your site
#
[tantek]
What is adding architecture tiers
#
Loqi
It looks like we don't have a page for "adding architecture tiers" yet. Would you like to create it? (Or just say "adding architecture tiers is ____", a sentence describing the term)
#
[LewisCowles]
I'll add one in a bit
#
[tantek]
adding architecture tiers is /architecture_astronomy
#
[LewisCowles]
Webmention on my site is the only thing that uses a database. I did not write it, nor do I consider it an anti-pattern
#
[tantek]
[LewisCowles] this "you'll hit file open limits if you don't make it into a pseudo database at some point" is demonstrably false. If you're curious about my setup, see /Falcon
#
[LewisCowles]
how are you avoiding file open limits which every single OS ever made has either explicitly (planned for failure) or implicitly (file handles exhaust ram).
#
[tantek]
I mean I had to invent a new calendaring system but it works 🙂
#
[tantek]
[LewisCowles] plenty of us are handling it just fine: https://indieweb.org/file-storage
#
[LewisCowles]
works on one property is an exercise in me either probing / attacking a system I have no sight of, or accepting the claim. I'd feel more comfortable with some form of what that means as it's a blurry picture to me. I am reading the article which defines an application-specific semi-opaque (with desires to be fully open) database. It might not be MySQL, but it's probably more rigid and less standard / open to comprehension than SQL
#
[LewisCowles]
Not an attack btw
#
[LewisCowles]
more... How does it mitigate file open limts?
#
[tantek]
SQL is more of a protocol (requires running software) than a file format (requires interoperable storage)
#
[tantek]
Maybe you're making architectural assumptions that would errantly hit file open limits?
#
[LewisCowles]
so many things have implemented SQL, including many "NoSQL" stores, that it's likely more interoperable than any OS filesystem, due to FS differences even within configurations of a FS
#
[tantek]
Unlikely, pretty historically well documented.
#
[LewisCowles]
Well the only assumption to counter is limiting the number of visitors
#
[tantek]
Check out the examples if you like to understand how lots of us have solved this, have it running in the wild just fine: https://indieweb.org/file-storage#IndieWeb_Examples
#
gRegorLove
On the plus side /database-antipattern reads much better than it used to several years ago :)
#
[LewisCowles]
can you see tantek how that is very xkcd to make a new format to replace all the flawed inconsistent ones out there?
#
[tantek]
Error establishing a database connection << database anti-pattern
#
[LewisCowles]
it's linked the other way, but locality++
#
Loqi
locality has 1 karma over the last year
#
[LewisCowles]
also better faith argument++
#
Loqi
argument has 1 karma over the last year
#
[tantek]
not inconsistent, fragile & unreliable
#
[tantek]
use database software, expect loss of data over time
#
[LewisCowles]
I mean supplant any property with another
[Aaron_Klemm] joined the channel
#
[LewisCowles]
you do have a solution in files to memory map your way through though. Falcon++
#
Loqi
Falcon has 1 karma over the last year
#
[LewisCowles]
for an individual website that's a nice touch
#
[Aaron_Klemm]
Looking at IndieAuth and wanting deploy my own authorization endpoint, can anyone catch me up A) the state of IndieAuth...all good and healthy and recommended? B) Which software should I use. self-auth, acquiescence, and cellar door look like the candidates. Which is considered ready to use and remains maintained?
#
[Aaron_Klemm]
I've read the excellent IndieAuth page on the wiki and looked through the activity on each project, but still am a little fuzzy how to proceed. I don't understand the flow as clearly as I did OpenID way back in the day when I last implemented such a thing.
#
Zegnat
I am a contributor to selfauth, and am also (for now) still actively running it myself to log in to IndieWeb services. So that one I know still works.
#
Zegnat
What is cellar door?
#
Loqi
It looks like we don't have a page for "cellar door" yet. Would you like to create it? (Or just say "cellar door is ____", a sentence describing the term)
#
[LewisCowles]
IndieAuth is great. I use IndieLogin.com (canonical). It's maintained
#
[Aaron_Klemm]
Cellar Door is a project that aims to provide a well-tested personal authorization server that implements the IndieAuth protocol. https://github.com/nilsnh/cellar-door
#
Loqi
[nilsnh] cellar-door: Personal authorization server implementing the IndieAuth federated login protocol.
#
[LewisCowles]
I've contributed towards it and have found that a valuable way to learn
#
Zegnat
But yeah, several of those solutions are still being maintained. Though I would be biased when recommending selfauth ;) I do find that because we intentionally kept the code short, it might be a good jump off point if you want to learn about the implementation
#
[Aaron_Klemm]
Ok, selfauth sounds great. Thanks for validating!
leg, sivy and andysylvester joined the channel
#
@knuspermagier
Ich würd ja Webmentions in den Blog einbauen, aber wo bringt man da die DSGVO-Checkbox unter.
(twitter.com/_/status/1247252326819213315)
[jeremycherfas] joined the channel
#
[tantek]
Welcome [Aaron_Klemm]++
#
Loqi
[Aaron_Klemm] has 1 karma over the last year
#
[Aaron_Klemm]
Glad to be here
#
@mauricerenck
↩️ Ich habe ein Kommentarplugin für Kirby geschrieben, dass auch webmentions macht, aber daran habe ich tatsächlich nicht gedacht. Wenn man selber welche verschickt ist das ja kein Thema, aber ich hole mir auch Twitter replies rein, das ist könnte also wirklich relevant sein...
(twitter.com/_/status/1247268013771980800)
#
@RubygemsN
webmention (2.1.0): A Ruby gem for sending Webmention notifications. https://rubygems.org/gems/webmention
(twitter.com/_/status/1247271975224889345)
#
jgmac1106_
snarfed not critical at all but trying to determine if I could send a webmention from http://longthoughts.jgregorymcverry.com/doku.php?id=namespace:update_on_indieweb_wiki tells me to add a profile page but I can’t find the exact link
#
Loqi
namespace:update_on_indieweb_wiki
[snarfed] joined the channel
#
[snarfed]
jgmac1106 sorry, this is with bridgy?
#
jgmac1106_
yes, my mistake, and trying to figure out what link, if any possible to add to my twitter bio, its really for S&G, just peaked my curiosity
#
[snarfed]
ah. bridgy publish?
#
[snarfed]
i see http://longthoughts.jgregorymcverry.com/doku.php?id=start in your bridgy account. try dropping the path? ie put just https://longthoughts.jgregorymcverry.com/ in your twitter bio
#
Loqi
start
#
jgmac1106_
yes it tells me to add: y. Check that longthoughts.jgregorymcverry.com/doku.php?id=na.../about#profile-link">
#
jgmac1106_
okay will do thx
#
jgmac1106_
that worked I need to add a p-summary, thanks
#
jgmac1106_
I think in a weird way, this was so easy to do and with a blog and a wiki, people I know may want use a tool like this
#
[snarfed]
glad to hear it!
Giske, loicm and [Rose] joined the channel
#
[snarfed]
if i want to find an h-feed's "author," i'm currently using https://indieweb.org/authorship, but i suspect that's wrong, since it's meant for posts. sounds like i should use the h-feed's p-author mf2 property, and if that's not provided, fall back to... https://indieweb.org/representative_h-card ?
#
[snarfed]
happy to ask in #microformats instead if necessary
#
Loqi
[alexmingoia] #195 Incorrect author identification for HTML source
#
sknebel
[snarfed]: I feel like the autorship algorithm (https://indieweb.org/authorship-spec) mostly applies. replace "h-entry" with "h-feed", skip step 4
#
sknebel
but the "turn a URL into an h-card" mechanism still applies I think?
#
sknebel
so a modified version of the algorithm makes sense IMHO
#
[snarfed]
hmm ok!
#
GWG
[snarfed]: I think I wrote something about this too
#
[snarfed]
GWG: um ok! please do find it if you can
#
GWG
On the /h-feed page, I made a list of sites that had author h-cards at feed level
#
GWG
It was when I was talking about implied h-feed and consistent data
#
GWG
[snarfed]: Have I told you that I've been following your lead of late?
geoffo joined the channel
#
[snarfed]
GWG you haven't!
#
[snarfed]
also GWG i'm sure you've heard it before, but you often make leading comments and ask leading questions without actually providing the obvious info you're alluding to. eg what you wrote about h-feed (above), what you're doing to follow my lead, etc
#
GWG
Actually the delay the first time was me eating dinner.
#
[snarfed]
consider providing that info up front, instead of waiting for people to ask you for it 😁
#
[snarfed]
ah ok! then i take back that one
#
[snarfed]
anyway! what lead?
#
GWG
Re the lead, bridging things a lot lately
#
GWG
I have been thinking about it since the newsblur to Microsub system you did
#
GWG
I keep looking for more things to throw in.
#
GWG
I have been going through all the sources I use regularly
#
@ChrisAldrich
↩️ I wrote and intro article a while back for A List Apart and @JmacDotOrg just wrote a good one. Perhaps @JGarber might be a useful resource within the Ruby world? Next take a look at the Webmention page of the IndieWeb wiki which has resources [more...] https://boffosocko.com/2020/04/06/webmention-and-ruby-related-resources/
(twitter.com/_/status/1247293606777389056)
#
sivy
hey folks
#
sivy
dumb q
#
[snarfed]
hi sivy
#
sivy
any idea why I can reach an amzn.to link from my worksatation fine, but from my linode its gets a 503
#
sivy
evn weirder, with curl on the linode i get a 301
#
sivy
locally i get 200 (probably post-redirect)
#
sivy
let me check somethng else
#
sivy
yeah if i load the redirected url from the linode it’s a 503
#
sivy
(basically this is in a webmention context)
#
sivy
hi [snarfed]
#
jmac
Hmm, just discovered that sometime between midday yesterday and now, events.indieweb.org reclassified my RSVP (from my big numb blog post yesterday) as pending, and removed it from the specific event page
#
jmac
Would it do that if I re-sent the RSVP?
#
jmac
Coz my blog 100% did that, a bunch of times, every time I moved a comma in the post
geoffo_, TGiske, geoffo, Kaja and Giske joined the channel
#
[LewisCowles]
sivy, the 301 is the redirect
#
[LewisCowles]
you may need to preserve some state like cookies
#
[LewisCowles]
browsers do this by default. CURL may not, which says awful things about t.co links, but it's equally as likely the endpoint in question is genuinely returning a 503 sometimes
#
[LewisCowles]
also apologies. amzn.to not t.co although I like neither
#
aaronpk
sivy: amzn.to is likely doing user agent sniffing and returning different results depending on what it thinks is requesting it
#
aaronpk
try adding a browser user agent to your code that's getting a 503
#
aaronpk
jmac: it puts it back in the queue if anything has changed, which is the paranoid thing to do, to prevent someone from updating a response with changed content after it's been approved
#
aaronpk
but again if you just log in to the wiki with that domain it will stop moderating those webmentions
[KevinMarks] joined the channel