#dev 2018-04-09

2018-04-09 UTC
jjuran and eli_oat joined the channel
#
GWG
I hope it is
#
grant.codes
edited /Micropub-extensions (+504) "/* Visibility */ Add thoughts on posts hidden from a main feed"
(view diff)
tantek, renem, eli_oat and dougbeal|mb1 joined the channel
#
GWG
I think it was a good week
j12t_, renem and [snarfed] joined the channel
#
Loqi
gwg has 27 karma in this channel (326 overall)
#
GWG
[snarfed]: I forgot that your push script only has two files and there are now three
KartikPrabhu, ben_thatmustbeme, jalcine and loicm joined the channel
#
aaronpk
grantcodes: I don’t think I’ve heard of that called anything before
#
aaronpk
i know several people have certain kinds of posts not appear in their main feed, I just don’t think I’ve heard of a name for that behavior
barpthewire and cweiske joined the channel
#
Zegnat
grantcodes, I have a checkbox “Sow in feeds” in my editor: https://licit.li/edit
#
Zegnat
s/Sow/Show/
[kevinmarks] joined the channel
#
[kevinmarks]
There is nipsa, but that's a bit different. What does mastodon call it?
AngeloGladding joined the channel
#
Zegnat
Hmm, I think I have seen unlisted used before. Makes sense as a term: the post is published and available at its URL, just not shown in any index/feed/overview
swentel joined the channel
#
swentel
Loqi, tell aaronpk still no updates on my timeline in aperture - but no rush, I can always delete and reimport if needed :)
#
swentel
!tell aaronpk still no updates on my timeline in aperture - but no rush, I can always delete and reimport if needed :)
#
Loqi
Ok, I'll tell them that when I see them next
#
swentel
ok, different way to tell heh
[pfefferle] joined the channel
#
@MorganGeek
↩️ @hsablonniere Connaissais pas Webmention :) , intéressant !
(twitter.com/_/status/983288735247880192)
[kevinmarks], [unoabraham], jeremycherfas, [jgmac1106] and loicm joined the channel
#
@Inkwater_Masha
Publishing is a choice. Here's an interesting perspective of the challenges and reality out there. http://ow.ly/23Qt30jjmqs #publishing #self-publish #traditional #amwriting #indieauth #author
(twitter.com/_/status/983329980942823426)
#
@kevinmarks
↩️ microformats are still a thing - the microformats2 revision simplified them, and they are being used as part of the #indieweb cross-site communication with webmentions etc. https://indieweb.org/posts#Types_of_Posts - I compared with schema before http://www.kevinmarks.com/microformatschema.html
(twitter.com/_/status/983330001729851392)
[pfefferle], singpolyma, leg, tantek_ and [kevinmarks] joined the channel
#
tantek_
Wow in the first four days of April, this IP address did 300k requests for ~4GB from my site: 194.74.171.82 - supposedly in Liverpool, England. Y THO? Anyone else see that address in their logs?
#
aaronpk
wow xmlrpc really complicates things for no good reason
#
Loqi
aaronpk: swentel left you a message 6 hours, 2 minutes ago: still no updates on my timeline in aperture - but no rush, I can always delete and reimport if needed :)
[davidmead] joined the channel
#
aaronpk
but I now have an XMLRPC->Micropub proxy
#
aaronpk
just functional enough to post content, only proxies the `newPost` function
[jgmac1106] joined the channel
#
Loqi
[Aaron Parecki] A MetaWeblog to Micropub Gateway
#
aaronparecki.com
edited /MetaWeblog (+266) "micropub gateway"
(view diff)
#
gRegorLove
whoa, awesome!
tbbrown, singpolyma, snarfed, [kevinmarks], KartikPrabhu and fdfsd joined the channel
#
skippy
that's very cool.
[cleverdevil] joined the channel
#
Loqi
aaronpk has 133 karma in this channel (1614 overall)
#
[cleverdevil]
I started working on one of those about a year ago and ran out of steam 🙂
#
[cleverdevil]
Interested in trying this for MarsEdit
#
aaronpk
I know MarsEdit implements way more of the metaweblog API so i'm curious how much of it it actually requires
#
[kevinmarks]
XMLRPC has C calling conventions so no named parameters
#
[kevinmarks]
aaronpk++ nicely done
#
Loqi
aaronpk has 134 karma in this channel (1615 overall)
#
aaronpk
[cleverdevil]: haha yeah, last time i tried this i also ran out of steam
barpthewire, [jgmac1106] and KartikPrabhu joined the channel
#
Zegnat
aaronpk, is XRay getting https://grapefruit.zegnat.net/#dt201806091719Z wrong? Shouldn’t it report the URL of the entry after resolving it based on the <base> element?
#
Loqi
Have we officially left nice emails behind? I usually use an “interleaved posting” style when writing email. Today someone ended an email to me as follows, and my heart broke a little: Btw, your replies are ...
#
aaronpk
e-item??
#
Zegnat
succeeds in edge cases even when handrolling his HTML 🎉
#
aaronpk
i don't think i have any <base> support in XRay right now
#
Zegnat
Oh, yeah, the e-item is there because of the h-feed and most of the items not having any other mf2 on them.
#
Zegnat
So theoretically the page is still a valid feed.
#
Zegnat
But XRay correctly gets the h-entry for the fragment link. But looks like it doesn’t resolve the u-url
#
Zegnat
Hmm, I should probably change that e-item to e-entry, per proposed properties on the mf wiki
#
Zegnat
(Becaus e-entry isn’t at all going to get confused with h-entry :P )
#
aaronpk
i don't understand
#
aaronpk
why not just use h-entry
#
aaronpk
nothing is going to consume e-entry or e-item
#
Zegnat
I am still messing about with the h-entry markup. That’s why only the most recent post has it.
#
aaronpk
those should be child h-entrys of the h-feed otherwise they look like properties on the h-feed
#
Zegnat
And I am using the e-item myself, in my little preprocessor, to walk the feed.
#
Zegnat
I couldn’t walk the items without putting some sort of property on them. Will move away from it once I decide on how I want my h-entry markup
singpolyma, snarfed and klandwehr joined the channel
#
realize.be
edited /Indigenous_for_Android (+0) "fix link to ios"
(view diff)
#
aaronpk
https://techcrunch.com/2018/04/07/rss-is-undead/ interesting thoughts about the problems with RSS feeds from a publisher/writer perspective
#
Loqi
ok, I added "https://techcrunch.com/2018/04/07/rss-is-undead/" to the "See Also" section of /RSS https://indieweb.org/wiki/index.php?diff=46743&oldid=46696
#
singpolyma
"they [users] want a feed or stream that shows them the most important content first" -- I often want this, by my anecdotal experience with less-technical users is that they *hate* this. I expect it's a matter of personal taste and should say "some users" :P
#
singpolyma
I mean, I literally worked at a company that provided this for feeds across the web, so certainly *someone* wants it :)
#
skippy
following more than about a dozen active content creators makes chronological following hard: you feel like you're missing too much.
#
skippy
and the constant drive to get more followers and "engagement" compounds the problem.
#
skippy
especially when it's all one feed. Twitter Lists, at least, let you carve up your feed
#
singpolyma
Maybe that's what it is. Power users who follow lots of publishers want filters and ranking, people who just follow their friends and parents keep hitting the "most recent" button on their facebook homepage
#
schmarty
I think with /Microsub we'll be able to see this with indie readers, soon.
#
schmarty
the ability to collect a bunch of feeds as "raw" data that a personalized service can rank and re-prioritize for viewing in an app like /Monocle or /Together or /Indigenous.
#
dgold
aaronpk++ for the metaweblog awesomeness
#
Loqi
aaronpk has 135 karma in this channel (1617 overall)
#
singpolyma
"Brands want users to know that the brand wrote something" and yet many brands happily publish on Facebook pages
#
singpolyma
schmarty: it depends how you want to do your rankings
#
singpolyma
at Postrank we had a *lot* of computers working very hard to collect the data needed to do what we did
[cleverdevil] joined the channel
#
[cleverdevil]
+1 on that [schmarty]... I'd really like to see a third-party IndieWeb version of something like Nuzzel.
#
[cleverdevil]
what is Nuzzel?
#
Loqi
Nuzzel is a service with an app that surfaces articles based upon how many of your friends like those articles https://indieweb.org/Nuzzel
#
singpolyma
Yeah, they look like they're using social engagement and such too, though maybe in a smaller way than we were (just looking at your contacts or something? hard to tell from the marketing)
[snarfed] joined the channel
#
[snarfed]
we should get andy to open source http://belong.io
#
aaronpk
oh yeah he said that's going to stop working too when twitter shuts down their streaming api
#
[cleverdevil]
I sort of like the idea of publishing an XFN list of people I "follow" on my website, and then having the third party service look at what they repost/like/bookmark, then publish things via Micropub to Aperture for discovery.
[davidmead] joined the channel
#
[davidmead]
@singpolyma - big issue I have with that is this…who decides what the “most important content” is?
#
[davidmead]
I am a die hard chronological order fanboy
#
skippy
just because X number of my friends engaged with something isn't a reliable indicator that I want to engage with it. I dont want to live in an echo chamber.
#
aaronpk
the whole point of the architecture of Microsub is that it turns out everyone can have different opinions on this and still use the same tools
#
singpolyma
[cleverdevil]: is that much different from just following your friends and reading what they post/like/share?
#
jalcine
yo aaronpk that XML-RPC bridge is _nice_. I've been looking more into making native clients for blogging more friendly - this looks like it
#
aaronpk
jalcine: thx! did you try it out? I'm a little scared of finding out how many more XML-RPC edge cases I'm gonna need to handle to support more clients
#
jalcine
I haven't yet - I don't actually have any that'd support it, lol.
#
jalcine
I might try it on my wordpress blog, just to test it out (haven't used that in ages tho)
tantek_ joined the channel
#
loqi.me
created /Personal_Democracy_Forum (+317) "prompted by tantek_ and dfn added by tantek_"
(view diff)
#
loqi.me
edited /Personal_Democracy_Forum (+55) "tantek_ added "2018: https://www.pdf-18.com/about/" to "See Also""
(view diff)
#
gRegorLove
One of my favorite Google Reader features was the ability to share items with your GR friends, then they could comment on it.
#
gRegorLove
The Nuzzel question reminded me of that.
#
gRegorLove
It was like a mini-subreddit for your friends, in the GR silo of course.
[kevinmarks] joined the channel
#
[kevinmarks]
Nuzzel uses your follows to rank links (and shows you who shared them)
#
[kevinmarks]
If you don't follow enough people to generate a good list of links, it goes out one more degree to your follows' follows
#
[kevinmarks]
It will also give you most shared links for your lists
singpolyma joined the channel
#
Zegnat
Alright, wrote a 65 line static site generator. So now https://grapefruit.zegnat.net/ should have a good ol’ h-feed :D
#
aaronpk
follows in aperture
#
tantek.com
edited /POSE (+428) "2007 early reference to concept, citation, quote"
(view diff)
#
aaronpk
hm, no author info showing up
#
Zegnat
I haven’t checked if I am doing authorship right yet.
#
Zegnat
The h-entry has a parent h-feed with an author property, so it should work if you have authorship implemented. Maybe.
#
aaronpk
i wonder if i just didn't implement that in xray? i see no author on the h-entry, so the next step says to check for author on the parent h-feed, which there is, and it's a URL to https://vanderven.se/martijn/
#
Loqi
Martijn van der Ven
snarfed joined the channel
#
aaronpk
weird i thought i had tests for all the authorship cases in there
#
Zegnat
This might be another edge-case. If you are parsing only the element refered to by the fragment in the mf2 parser, you are never parsing the h-feed?
#
aaronpk
hmm, but aperture parses the page as a whole
#
Zegnat
Right, because it is fetching the feed not the separate u-urls. Then it must not be implemented.
#
Zegnat
For the curious, the 65 line static site generator: https://gist.github.com/Zegnat/5abdf00989a4591ade74a33c41cf4dc5
kaushalmodi joined the channel
#
aaronpk
huh yeah i think somehow i missed that case
#
Zegnat
kaushalmodi, if you want to see what sort of DOM stuff I would have liked Hugo to do, see the gist ^^^ Though it may be slightly undecipherable.
#
Zegnat
I am happy to once again be the edge-case, aaronpk ;)
renem joined the channel
#
Zegnat
made another post and hopes it will pop up in aaronpk’s reader
#
aaronpk
it did!
#
Zegnat
Now if I could make the build happen on the server, I would be really pleased with myself. But that might be a tomorrow project.
#
tantek.com
edited /business-brainstorming (-110) "replace stub dfn to redirect to a broader page with examples"
(view diff)
AngeloGladding joined the channel
#
tantek.com
edited /social-business (+39) "see also cross-enterprise social business"
(view diff)
[jgmac1106] joined the channel
#
tantek.com
created /business (+35) "r"
(view diff)
#
jgregorymcverry.com
edited /private_group (+166) "/* Perspective and Itches */"
(view diff)
#
tantek.com
edited /business-models (+19) "linky, see also"
(view diff)
#
tantek.com
edited /social-business (+74) "linky linky, dfn needs fixup"
(view diff)
#
tantek.com
edited /IndieWeb_for_business (+202) "see alsos, emojicon, sessions"
(view diff)
#
tantek.com
edited /block (+17) "emojicon"
(view diff)
#
tantek.com
created /⛔ (+19) "r"
(view diff)
#
tantek.com
created /🏢 (+35) "r"
(view diff)
#
gRegorLove
Zegnat++ nice work on that static site generator
#
Loqi
zegnat has 52 karma in this channel (193 overall)
#
tantek.com
edited /downtime (+196) "emojicon, silo example"
(view diff)
#
dgold
zegnat's grapefruit doesn't show unread status in my aperture
#
Zegnat
1. Upload a single HTML file blog in IndieWeb style.
#
Zegnat
2. Break lots of things because who would ever do such a thing!
#
skippy
#winning
#
dgold
aaronpk: pretty sure you don't do that followup in xray - this is the thing I spoke about waaaay back in December
#
aaronpk
followup?
#
tantek_
kevinmarks++ for user-centric reframing: https://mastodon.social/@kevinmarks/99800694724137318
#
Loqi
kevinmarks has 8 karma in this channel (305 overall)
#
Loqi
[Kevin Marks] @Gargron the other way of thinking about this is more federation. Webmentions are not like pingbacks, they're like mastodon notifications. How about framing the question in user terms: Would you like Mastodon to connect notifications to a broader ran...
#
tantek_
that needs to go into an FAQ or misconceptions section somewhere
#
tantek_
wow the entire webmention faq is dev-centric: https://indieweb.org/Webmention-faq
#
tantek.com
edited /Webmention-faq (+567) "separate user-centric vs dev FAQs, add Are Webmentions like Pingbacks inspired by kevinmarks answer https://mastodon.social/@kevinmarks/99800694724137318"
(view diff)
#
tantek.com
created /🚧 (+22) "r"
(view diff)
snarfed joined the channel
#
tantek_
aaronpk, how much information about the author of a post does a Microsub server cache?
#
aaronpk
it stores just the name/photo/url for the author of each post
#
aaronpk
stores it in the post, not a reference
#
aaronpk
there's no concept of an address book/contact list right now
#
tantek_
what if the h-card has more than name/photo/url?
#
aaronpk
it only grabs those properties
#
tantek_
why not all the properties of author h-cards?
#
aaronpk
if i was storing h-cards as h-cards i would do that
#
aaronpk
but it only stores enough about the post to show it in a reader, each author of each post is totally separate data right now
#
tantek_
no wait I never said anything about post. think followings / subs
#
aaronpk
there's no concept of h-cards in the following list right now
#
tantek_
when I enter a URL of a person to follow, how much of the h-card there does the Microsub server cache?
#
aaronpk
it just follows the URL you give it
#
tantek_
oh just the h-feed or h-entry s
#
tantek_
it doesn't check for a representative h-card at that URL?
#
aaronpk
that was my shortcut to actually get something shipped
#
aaronpk
nope, there's no concept of people
#
aaronpk
i realized that was adding a lot of complexity to the system that wasn't actually needed in order to collect and show lists of posts
#
tantek_
that seems like the difference between feeds and people
#
tantek_
sure, not needed for a mvp
#
tantek_
how are you getting the icon for a URL?
#
aaronpk
there is none
#
tantek_
to show in said lists of subscriptions
#
Loqi
[Kevin Marks] @Gargron @0x1C3B00DA yes they do - there is a pattern of citation of blog URLs to refer to the blogger that goes back to the dawn of blogs - it's a very natural thing to do I wrote about this about 10 years ago: http://epeus.blogspot.co.uk/2008/01/ur...
#
aaronpk
it only has author photo of the post
#
tantek_
seems like a pretty logical next step, from caching feed at URL, to caching feed + representative h-card (if any) at URL
#
[kevinmarks]
multi-author blogs make "blog url as person" a bit more complicated
#
tantek_
and would enable much nicer display in lists
#
Loqi
[aaronpk] #13 Indicating Item Source
#
aaronpk
a separate idea of the feed vs the posts within it
#
tantek_
and if it had the whole h-card, it would enable things like searching over your followers for who is in a particular city for example
#
aaronpk
which allows for things like following multi-author blogs or aggregators
#
aaronpk
none of that is needed to know where people you're following are though
#
tantek_
huh? that's a double-negative
#
tantek_
I'm just proposing one way it could work
#
aaronpk
an easier path for me to implement that is to look at the location of each h-entry aperture encounters
#
tantek_
since you're already fetching that page, keep it's representative h-card
#
Loqi
[Aaron Parecki] create meta-channel for showing the last known location of friends
#
aaronpk
that would also require less work on the part of the people i'm following, since their location would be updated automatically based on the contents of their posts instead of having to update their bio
#
tantek_
that's more like something to enable a (slower, more delayed) marauder's map
#
aaronpk
also i don't update my bio when i travel so it would just always show i'm in portland
#
tantek_
I wonder how many folks post their location in posts though - or don't because of privacy
#
singpolyma
wouldn't this kind of thing be better served by "location posts" (ala "check ins") instead of changing the location in your hcard?
#
aaronpk
singpolyma: yes that's what i was getting at, except not limited to checkins, since other posts can have locations too, like photos tagged at a location
#
Zegnat
Hmmmmmm, I do not have a live location in my h-card. Will have to think about that.
#
aaronpk
oh no haha
#
skippy
whats the canonical way to say "this is where I live" in an h-card?
#
Zegnat
I am using `p-adr h-adr`
#
aaronpk
huh, "location" is not a property in h-card
#
Zegnat
Then again, I use the same for my postal address. I don’t think there was a better way
#
Zegnat
I’ll make it one, aaronpk :P
#
Zegnat
skippy see https://vanderven.se/martijn/ for an example of displaying addresses
#
Loqi
Martijn van der Ven
#
tantek_
aaronpk - geo is the property you're looking for in h-car
KartikPrabhu joined the channel
#
tantek_
(The Geo was not a very good car, sorry)
#
aaronpk
geo has to be a lat/lng
#
tantek_
correct, and it's meaning is *actual* location, not something like an address (sometimes location)
#
tantek_
its meaning
#
aaronpk
hm yeah i think i'm gonna stick to my plan of extracting locations from h-entrys :) that seems to be much more consistent and widely implemented
#
tantek_
I'd be surprised if most people I followed posted location info in their posts
#
Zegnat
skippy, one of my addresses contains a post-office-box property, the other a street-address property. So theoretically you can programatically see which one of my addresses is postal and which is residential. But I don’t think anyone is doing anything like that.
#
skippy
how can i be less precise? i dont want to publish my home address.
#
Zegnat
leave the street-address property out
#
[kevinmarks]
just your locality?
#
aaronpk
plenty of my friends include the locations in their instagram and twitter posts
#
skippy
[kevinmarks]: yes; just my city.
#
Zegnat
h-adr with only p-locality and p-country-name could do, skippy
#
[kevinmarks]
it's funnty that those are different things in uk vs US
#
tantek_
aaronpk I agree that the post location info is more likely to be updated than bio / profile
#
skippy
and bio/profile is different from representative h-card? still confused.
#
[kevinmarks]
US "city" is often fairly small
#
[kevinmarks]
UK City is a big thing, with a royal charter
#
[kevinmarks]
We also have postal towns which are big areas that don't always correspond to historic boundaries , which can mess things up
#
skippy
so... p-locality and p-country-name both get nested inside p-adr h-adr? just making absolutely sure. v
#
Loqi
h-card
#
Zegnat
skippy, yes
#
Zegnat
have a look at https://vanderven.se/martijn/ if you want an example :)
#
Loqi
Martijn van der Ven
#
Zegnat
Examples of lots of different h-card things there
#
skippy
missed that before. thanks Zegnat
#
tantek_
no you don't need to nest them
#
skippy
but i have my h-card on my home page, rather than on a specific bio page.
#
Zegnat
That’s fine, skippy
#
skippy
tantek_: how would that look, un-nested?
#
tantek_
only nest them if you want to show / represent multiple addresses with that information clustered per address, which is not the typical bio page or profile
#
tantek_
skippy like any other microformats property, inside the h-card
#
Zegnat
Right, you can do that too. Guess I just prefer having the nested object for it always. Not sure why.
#
tantek_
Zegnat, unless you're showing more than one address (e.g. home vs work), then it's extraneous markup
#
tantek_
it was required to be nested in classic hCard, perhaps that's why?
#
tantek_
that was an intentional h-card simplification for like the 99% case
#
Zegnat
I am showing multiple adresses, so I would do it anyway. It might just be that I am so used to abstracting data to separate objects and addresses make perfect sense for that.
#
skippy
does this look right?
#
skippy
<div class="p-locality">Columbus</div>, <div class="p-region">Ohio</div>, <div class="p-country-name">USA</div>
#
skippy
i guess those should be spans, not divs?
#
skippy
<-- not a front-end dev.
#
Zegnat
spans are inline, divs are block, usually. So yes, spans would be nice :)
#
Zegnat
Otherwise, looks right to me!
#
Loqi
Scott Merrill
#
Zegnat
Nice!
#
tantek_
skippy - for microformats purposes, either of those tags work fine. Use the most appropriate HTML tag for the semantics in the markup. in this case div vs span is basically equivalent and you can pick based on what works in the context.
#
tantek_
E.g. inside <p> tag you'd have to use span because you can't put divs inside p
#
skippy
thank you!
snarfed, [kevinmarks] and tantek_ joined the channel
#
tantek_
how not to POSSE, or how to piss-off Kevinmarks about images of text: https://www.instagram.com/overheardmarinasf/ (which for some reason is not on Twitter? their IG is literally what look like images of tweets!))
#
[kevinmarks]
looks like python xmlrpclib is not good at showing errors
#
tantek.com
edited /search (+389) "emojicon, Twitter 2011 screenshots"
(view diff)
#
Loqi
xmlrpc has -1 karma in this channel (-3 overall)
#
loqi.me
created /bridge_all_the_things (+292) "prompted by snarfed and dfn added by snarfed"
(view diff)
#
loqi.me
edited /bridge_all_the_things (+87) "[kevinmarks] added "http://epeus.blogspot.com/2003/09/how-to-atomize-or-de-atomize.html" to "See Also""
(view diff)
[snarfed] joined the channel