#dev 2018-09-18

2018-09-18 UTC
[jgmac1106], iasai, [kevinmarks] and jjuran joined the channel
#
[jgmac1106]
and I lost all my screenshots when my hard drive crashed laways felt so ephermal...note to world need a screenshot micropub client
#
boffosocko.com
edited /disclosure (+1145) "Brainstorming: disclosure page for potential research conflicts in academia"
(view diff)
#
Loqi
I agree
#
[jgmac1106]
We can say over 1 million webmentions but thats too big...if folks have a comments or webmention counter and posted a screenshot, i would make text data as well for readers and parsers, and make a collage or gif...to answer question "Comments are Dead"
#
tantek__
[jgmac1106]: might depend on the context of said "Comments are Dead"
#
tantek__
... comment
#
[jgmac1106]
like in that tweet, as a rebuttal, well no here is actual data that they are alive and well
#
[jgmac1106]
unless we want to call webmentions undead....totally cool with that as well
#
[jgmac1106]
worry zom,bies may hurt D&I efforts
#
[jgmac1106]
nevermind just one, all ht elinks to my old bookmarks are dead
#
tantek__
[jgmac1106]: "Comments are dead in large swaths of the blogosphere" is a reasonably accurate statement though, hence why I asked about the context
#
tantek__
could be an opportunity to point out half-empty / half-full perspectives that can both be true
#
[jgmac1106]
Yes, that makes sense, and for many good reasons
#
tantek__
specifically with blog-to-blog comments *growing*
#
[jgmac1106]
good point it was when large publishers added commenting that a lot of the well got poisoned
#
tantek__
I tend to agree with some of the sentiment of that tweet in that *native* blog comments are pretty much dead IMO
#
[jgmac1106]
I guess this is where its so different coming from education...it never really stopped...leaves me with a bias opinion in many ways
#
tantek__
I used to debate this with benwerd (who still believed, may still believe) in native comments (which Known supports), and he had them enabled on his blog werd.io
#
tantek__
then fast forward a few years later, and vast majority of such native comments *on his blog* were just spam (manually entered)
#
tantek__
so I think he has terminated them
#
[jgmac1106]
Probably again from the education market, commenting is very strong but niche corner
#
tantek__
[jgmac1106]: I'm not sure what you mean by "different coming from education...it never really stopped" with respect to comments, do you mean in a "behind a firewall" like situation in colleges / enterprise?
#
[jgmac1106]
thinking I can actually demonstrate this by determining when I really enabled webmentions
#
[jgmac1106]
no really the small tight nit open community and all of edutwitter as a whole....I am sure twitter ate up a lot of blog commenting
#
[jgmac1106]
...actually scrolling through the top 20 blogs here...and clicking on random post I see no comments: https://teach.com/what/teachers-know/teach100/
#
[jgmac1106]
I mean they are ranked by social media not their comment feeds
#
[jgmac1106]
Now I do want to quantify this, check the last five post of everyone in my following page,
#
[jgmac1106]
clearly those with webmentions will have more comments, but then I could prove to myself that yes besides the people in my #openpedagogy feeds comments are dead
#
[jgmac1106]
...this isn't dev sorry...I could then find the exact moment when I moved from Known to WordPress and figured out when I turned webmentions on I can show a dramatic spkie in comments
renem joined the channel
#
jgregorymcverry.com
edited /disclosure (+67) "/* Examples */"
(view diff)
#
jgregorymcverry.com
edited /disclosure (+97) "/* Examples */ added example"
(view diff)
#
loqi.me
created /social_feed (+19) "prompted by tantek__ and redirect added by tantek__"
(view diff)
leg and dparrish joined the channel
#
dparrish
Alⅼah is doing
#
www.christopheducamp.com
created /indieweb.me (+3168) "First alpha description, stub, stub. just to let you know what I have in my head for the indieweb.me roadmap : a dream of indiewebFarm powered by Masto"
(view diff)
#
boffosocko.com
edited /disclosure (+79) "/* Brainstorming */ Jonathan Eisen example"
(view diff)
iasai and jjuran joined the channel
#
www.christopheducamp.com
edited /indieweb.me (-1) "/* = Mastodon */"
(view diff)
tantek joined the channel
#
Loqi
xtof has 1 karma in this channel over the last year (2 in all channels)
#
jeremycherfas
Re: multiple feeds. I find it really handy that WithKnown has a feed for each type of content, for example to feed micro.blog selected stuff only. But maybe tantek__ you would consider that personal.
electrobot17, cweiske, jjuran and tantek joined the channel
#
tantek__
jeremycherfas: it seems like extra complexity that feels like more "work" to setup / configure / use etc.
#
tantek__
the closest thing to "multiple feeds" that I've seen that seems both convenient (for publisher) and somewhat predictable / usable is (hash)tag based, e.g. what aaronpk does with his tag pages
#
tantek__
instead of by "type" of content since stuff can often be multiple types or types you never thought of.
Ruxton, swentel, tabaqui1 and jeremych_ joined the channel
#
jeremycherfas
don't personally explain that well on my site.
#
jeremycherfas
Withknown has defined types (post, photo, status, location etc) so types never thought of is not currently an issue. It can also feed specific tags, so that's a non issue too. And for someone who doesn't want to spray all their content unthinkingly into, say, microblog the extra minor complexity is definitely worth it. Can be handy for people who only want to receive certain sorts of content too, although I
#
cweiske
so what was your shpub bookmark html command that failed?
#
Loqi
cweiske: jeremycherfas left you a message 4 days, 14 hours ago: I'm just stopping for supper. Can I pick this up again tomorrow?
#
jeremycherfas
Will get back into it in about an hour. Will you be around?
#
cweiske
till 14:20
#
jeremycherfas
Cool. Thanks.
#
jeremycherfas.net
edited /RSS (+128) "/* Praise */ Add Brent Simmons piece"
(view diff)
thalios26 and bb1 joined the channel
#
@jgmac1106
↩️ Piece was great @zoecorbyn but think decentralized web imore a human problem than a technical one. We need to place users and not large social media in charge. Blog again. Encourage newspapers to send/receive webmentions, hint hint @guardian (http://jgregorymcverry.com/7025-2/)
(twitter.com/_/status/1041993023998099456)
barpthewire joined the channel
sophiebits joined the channel
#
sophiebits
Allaһ is doinɡ
[jgmac1106] joined the channel
#
www.christopheducamp.com
edited /indieweb.me (+548) "updating mission statement and personal research - delegation & governance"
(view diff)
#
aaronpk
I like my hybrid approach
#
aaronpk
my different feeds are not strictly based on type
#
aaronpk
but i have rules that sort posts based on type into the different feeds
#
aaronpk
but i can always override them
#
aaronpk
and it happens automatically unless i want to specifically change one
#
[jgmac1106]
that is neat, I kinda did the same making a feed called "stream" whihc is a mix of multiple post types and one "article" of course folks can subscribe to any post kind as a feed
#
[jgmac1106]
I also created a specific feed for micro.blog so now I can be really intentional about how I pPoSSE between Twitter, Mastodon, LinkedIn, and Micro.blog
[schmarty] and [eddie] joined the channel
#
aaronpk
i have just aaronparecki.com/notes and aaronparecki.com/photos going to micro.blog, so you see slightly different stuff there than my home page, but i haven't actually special-cased micro.blog yet
#
aaronpk
i don't think of it as syndication i guess
#
[eddie]
I had micro.blog as a special case syndication, but in my recent incarnation, I have a manual channel on my site called Timeline and whatever I tag as being in the Timeline channel appears on Micro.blog. So it won't be special cased anymore but it will still be manual because only things I manually tag will appear in the Timeline
#
[jgmac1106]
is this how folks would correctly maerk up a follow post? http://jgregorymcverry.com/7001-2/
#
Loqi
[Greg McVerry] I just added @tomcritchlow’s Writing Page to the #BloggersILike section of my following page https://jgregorymcverry.com/following #BeYourOwnSocialNetwork
#
aaronpk
oops there's an extra " in there messing up the html
#
aaronpk
i think the current idea is to use "u-follow-of" https://indieweb.org/follow#Indieweb_examples
#
aaronpk
otherwise that looks great! good fallback text that will look good on your site and in readers
#
[jgmac1106]
I know [tantek] doesn't like to display code on post types considered experimental pages but I find it super helpful to experiment when the examples include a how to markup: https://indieweb.org/event
#
[jgmac1106]
maybewe could encourage folks that do put examples on experimental post types
#
[jgmac1106]
to include a how-to markup of their example so folks can compare
#
[jgmac1106]
then if it is an accepted format the markup gets moved to the top of the article
#
swentel
aaronpk, I guess xray already supports that functionality (getting content from a site), so getting the reply/like context probably does belong in aperture
#
GWG
Aaronpk, I am interested in your ideas for xray due my similar product
#
swentel
might be an option in aperture in channels - get contexts from those types maybe as it's more processing
#
[jgmac1106]
the fact that I am using u-follow-of in a note is okay?
#
aaronpk
[jgmac1106]: yeah a follow post shares a lot of properties with h-entry (published date, author, etc) so it makes sense to use that
#
[jgmac1106]
chris created a post kind it looks like that uses <section class="response u-follow-of h-cite">
#
aaronpk
swentel: yeah i'm on the fence about it. right now if someone's post contains a nested h-entry in the "in-reply-to" property, xray does include the external content but not fetched from the URL, just from the embedded markup
#
aaronpk
and xray does already make up to 1 additional request to go fetch the author h-card
#
aaronpk
i'm just a little worried about having xray start making potentially a bunch of additional requests that the thing calling xray may not be aware of
#
aaronpk
but it's probably just a matter of making sure people are aware that's happening if they use it, so maybe making it an opt-in or opt-out feature
#
Zegnat
[jgmac1106], the h-cite being on the same element as u-follow-of means that it is nested under the follow-of property. Probably as a way to put both a name of a person and their url in the mf2 (just u-follow-of will only result in a url).
#
aaronpk
i think i would need to also include some information about whether a particular object was fetched from the source or not, for security reasons
#
Zegnat
Chances are that code from chris is still within an h-entry
#
GWG
aaronpk, how does it decide to fetch an external h-card?
#
aaronpk
GWG: it follows the /authorship algorithm which says to fetch the author page at some point
#
[jgmac1106]
<a href="http://catherinecronin.net/" class="p-name u-url">Catherine Cronin – open educator, open researcher, educational developer</a> <em>(<span class="p-publication">catherinecronin.net</span>)</em></header>
#
[jgmac1106]
<blockquote class="e-summary"><p>I’m Catherine Cronin — open educator, open researcher and educational developer in CELT (Centre for Excellence in Learning &amp; Teaching) at the <a href="http://www.nuigalway.ie/" target="_blank" rel="noopener noreferrer"
#
GWG
aaronpk, I am worried about multiple calls and am thinking about having that as optional
#
aaronpk
GWG: I would be more worried about it in your case since wordpress sites typically run everything in the same web request
#
aaronpk
in Aperture, at least those requests are all happening in a background queue
#
Zegnat
considers formally writing up his mf2 primer to do at gatherings to explain parsing
#
Zegnat
That might be able to explain things like why you might consider adding an h-cite as a property.
#
GWG
aaronpk, exactly. That's why I am working on it
#
aaronpk
Zegnat: do it! that sounds very useful
#
GWG
Why might I want a nested h-cite?
#
GWG
Asking for a friend
#
Zegnat
I think I did a bit of a primer as a session in Berlin. Got some questions that I hopefully answered in a way people understood. Turning it into a presentation could be interesting.
#
GWG
And why does jf2 have it in a new property over mf2 where I would nest it?
#
Zegnat
GWG: whenever it makes sense to provide more data than just a URL, an h-cite can often fill the gap. A u-like-of is nice when it is a URL, but it doesn’t say much to a reader. A u-like-of being an h-cite means you can provide a separate title and url for the like.
#
aaronpk
Zegnat: not a separate title for the like, but providing the title of the thing being liked
#
GWG
I do that. Often more than just name and title
#
aaronpk
an h-cite shouldn't include any new information that isn't already at the thing being cited
#
GWG
Author, featured image, etc
#
Zegnat
Yes, you are right aaronpk, sorry if I implied otherwise.
#
Zegnat
Using h-cite with title and url for a u-follow-of is the difference between saying “I am following https://david.shanske.com/series/indieweb-podcast/” and “I am following the IndieWeb Podcast at https://david.shanske.com/series/indieweb-podcast/”. When I read the mf2 output.
#
GWG
So why does jf2 have reference?
#
Zegnat
I am not really familiar with jf2
#
aaronpk
GWG: I wanted to keep things relatively flat in jf2, and avoid properties switching between strings and objects
#
aaronpk
so an "in-reply-to" value will *always* be a URL in jf2, and *if* in the mf2 that happens to be an h-cite, then the nested object gets put in "refs" keyed off of the URL so you can match it up if you want
#
aaronpk
the plan with actually fetching that page is that I can then also include the external content in the refs object as well
#
Zegnat
If I am reading this right it is recommended that references are always fetched externally also? That’s interesting because they would overwrite an h-cite that may include a specific summary
#
aaronpk
yeah i would overwrite the h-cite data from the external content
#
GWG
I am nesting h-cites in jf2 in my latest fixes. I will have to fix that in a future version
#
aaronpk
because an h-cite shouldn't have new information so in theory it should always be less useful than the info at the source
#
Zegnat
Interesting, so a quote post can’t exist in jf2?
#
GWG
I finally got my mf2 looking right
#
sknebel
it can, jf2 allows nested objects, it just does recommend against it
#
sknebel
e.g. the thing might not even have a url
#
GWG
I am using nested objects now, against recommendations
#
GWG
Too much to keep track of
#
Zegnat
I quote a paragraph, mark it up with all the right citations, and in jf2 my single paragraph may be overwritten with the entire contents of the external thing? In that case I would definitely stick with a nested and not a ref. Though that’s one very specific use-case.
#
aaronpk
looking at /quote it looks like that kind of post isn't even well defined yet so it's quite possible that "u-quote-of h-cite" is not the right solution for that
#
Zegnat
could be. Same with .u-follow-of.h-cite if that suddenly gets replaced with an entire h-feed object.
#
Zegnat
Too many different h-cite uses, he
jackjamieson joined the channel
#
aaronpk
this is why when i add this to aperture/xray i'm going to be very selective about what external content i fetch and use
#
sknebel
Zegnat: h-cite does not intend for you to put arbitrary content in it from what I see?
#
GWG
aaronpk, what content are you thinking?
#
Zegnat
I think you can pick your own summary for marking up a blockquote, sknebel?
#
aaronpk
right now it's pretty well established for like-of/repost-of/in-reply-to
KartikPrabhu joined the channel
#
aaronpk
those are also the ones that i most often encounter and wished i had the full content available
iasai joined the channel
#
GWG
Figuring out content blocks on non mf2 sites is a pain
#
sknebel
Zegnat: maybe? summary is not listed as a property
#
Zegnat
I might be mistaken on the blockquote thing. Not sure where I picked that up...
#
sknebel
and content is clear aobut being the entire content of the source
#
Zegnat
h-cite is also used for its original purpose: citation links ;) But that’s not as a property so shouldn’t run afoul of jf2 much
#
Zegnat
https://indieweb.org/h-cite#IndieWeb_Uses also lists “blockquotes”, which I don’t think refers to an entire repost.
#
Zegnat
Darn I wish I could remember where this practice came from!
#
sknebel
the quote posts I looked at had the blockquote outside the h-cite I think
barpthewire joined the channel
#
[jgmac1106]
sounds like a session idea for Oxford this weekend to be finished in NYC and published in Berlin
#
Zegnat
I am really going to miss out on a lot of stuff by not doing Berlin this year, ha
#
Zegnat
sknebel, https://boffosocko.com/kind/follow/ - all have a summary on the h-cite
[tantek] joined the channel
#
[tantek]
[jgmac1106] not opposed to documenting how to markup. Opposed to claiming a “how to” for something that is just an idea or has zero consumers.
#
[tantek]
Start with putting such documenting in a “brainstorming” section so you’re transparent about the reality of it being an experiment and not “real” yet.
#
[tantek]
Until you have (preferably multiple) publishers *and* consumers (different code for each), it’s not real, it’s an experiment.
#
[tantek]
Also why u-follow-of h-cite?
#
[jgmac1106]
more -meta....but that got me thinking about placement...if it is documents as accepted the "how-to" could be at the top
#
[jgmac1106]
if it is an example someone adds to a post type page for an experimental they should be encouraged to include the markup and just the link
#
[tantek]
“Accepted” is not enough as that sounds like political agreement rather than implementation and deployment
#
[tantek]
They should be encouraged to include it yes but *only* in the brainstorming section until multiple others publish & consume it
#
[jgmac1106]
the top list here: https://indieweb.org/posts#Types_of_Posts has the "how to" up top
#
[tantek]
Otherwise any such claim of “how to” is an exaggeration and misleading
#
sknebel
Zegnat: seems mostly automatically extracted from the source, so fine to refetch? + remember that it doesn't get automatically thrown away necessarily, so it can still be used if extracting from the source fails
#
[jgmac1106]
so differentiate between "how-to" and "here's how I do it"
#
Zegnat
“why u-follow-of h-cite?” - to provide a title (and/or other context), much as for bookmarks and likes
#
Zegnat
sknebel, yeah, I get that :)
#
[tantek]
Zegnat no that’s insufficient and likely misleading. Most of the time you follow a *person* not a post
#
[tantek]
“To provide a title” works with any embedded microformat. That’s just p-name
#
Zegnat
If you are following a person rather than a feed (note: my example was a specific h-feed for a podcast) you can do u-follow-of h-card. Pretty much the same thing.
#
Zegnat
I agree that h-card is better if you are making a follow post for a person, for sure
#
[tantek]
The markup example given above was obviously a person, not a feed
#
[tantek]
And no reason to duplicate the url as a “p-publication”.
#
GWG
I wouldn't count chrisaldrich as he sort of hacked his implementation. I intend to add official support
#
[tantek]
The Catherine example from [jgmac1106]
#
[jgmac1106]
yes I was wondering if I should be following a n h-card I would write for Catherine in the post
#
Zegnat
Oh, I wasn’t particularly commenting on that example other than trying to explain why one would nest h-cite on a property in general. The markup on that example (which is from chrisaldrich) has a few more questionable things *shrug*
#
[tantek]
1 no need for p-publication (already in u-url)
#
[tantek]
2 use p-note for someone’s info, not e-summary
#
[tantek]
An h-cite is a shorthand reference for a piece of content. That’s it. If the content itself is small or can be summarized with just a few properties (and more info at a u-url), then better to use that microformat directly instead of an h-cite
#
[tantek]
Otherwise you’re making the excess abstraction mistake
#
GWG
Zegnat: the code he used uses h-cite as a default.
#
[tantek]
Default for what?
#
[jgmac1106]
<span="class u-follow-of"<a class="h-card u-url p-name" href="https://david.shanske.com/">David Shanske</a></span>?
#
[jgmac1106]
oops on the quotations, getting me all day..
#
GWG
tantek, everything. I have been working on the new version which is more intelligent
#
[eddie]
jgmac1106 I believe u-follow-of would be on the a element
#
GWG
And follow is on the list for me
#
[jgmac1106]
cool anytime I can drop a span I am happy
#
GWG
Big list
#
[eddie]
or maybe the h-card should be on the span
#
[eddie]
I'm not 100% sure
#
[jgmac1106]
I think he means there is no default for experimental posts
#
[jgmac1106]
yeah I didn't think they could be together...no real reason...just seemed like it would mess things up
#
Zegnat
<a class="u-follow-of h-card" href="$url">$name</a> - should be enough for following a person
#
[jgmac1106]
and what is rel=follow again? 😁
#
Zegnat
I don’t think there is a rel-follow?
#
[jgmac1106]
perfect @zegnat going on my stickies...which are now just manual mf2 tempaltes
#
Zegnat
Also, I’d feel rel-follow might point at the inverse of rel-nofollow. But rel-nofollow means a specific thing already that has nothing to do with our use of the word following.
#
Zegnat
Stick to the microformats, a lot more clear than rel values at times ;)
#
[jgmac1106]
I know, thus the smiley face, shoudl have used something a little more mischievous. i am sure there is something that dives deep into the history here: https://epeus.blogspot.com
#
Zegnat
If there is, give it a couple hours for kevinmarks to pop in and link it :D
Zathrus28, [kevinmarks], barpthewire and jackjamieson joined the channel
#
[kevinmarks]
I think we were converging on rel="following"
#
[kevinmarks]
Vote links was my attempt, which nofollow displaced.
#
[kevinmarks]
I still want rel="not-me"
#
Zegnat
For all those websites that do “for the other Martijn click here”?
#
[kevinmarks]
Yes, and (originally) for all the ones that copied my rel=me example too literally
#
petermolnar
ben.thatmustbe.us
jillr23 and [tantek] joined the channel
#
[tantek]
Kevinmarks, good reminder to use example.com urls in visible markup examples that people are likely to copy/paste
#
[tantek]
Zegnat that’s a decent brainstorm - simple one element markup - can you add to wiki brainstorming on follow page?
[schmarty], jgmac1106, barpthewire, Vlad11, jackjamieson, trippeh_9, KartikPrabhu, doubleloop, calumryan, [tantek] and tantek__ joined the channel
#
tantek__
tl;dr: Color is not used as the only visual means. That's sufficient. stressing any more about that may be a waste of time and/or an unnecessary sacrifice to otherwise explicitly designed colos for purely decorative purposes.
#
tantek__
s/colos/colors
a--m4 and dougbeal|mb1 joined the channel
tantek__ joined the channel