#dev 2018-01-25

2018-01-25 UTC
#
tantek
"Professional driver. Closed course."
#
tantek
new cassis and falcon deployed
#
tantek
looks like the new math is working
#
tantek
going for 0
leg joined the channel
#
tantek
well since I'm now *always* going to include a PSL
#
tantek
and tweets are longer
#
tantek
I'm putting the PSL on its own line
#
tantek
might as well throw a design change in the mix too
#
Loqi
[Tantek Çelik] #Twitter #developers heads-up:@Twitter just changed their backend & display auto-linkers (but not posting UI!) to link pathless ccTLD domains, when previously it did not.The API counts ttk.me as 23 (vs 6 before),but tweet UI does not!
#
aaronpk
:sigh: whitespace is hard
#
aaronpk
makes sense to me
#
tantek
posted with 0 chars remaining
#
tantek
and cassis.js updated on master with a fix to tw_text_proxy
#
tantek
and thus tw_length_check
#
tantek
in case anyone was using those and wants to update for Twitter's update
#
tantek
Well that's it for "This week in keeping up with random silo API changes"
#
tantek
aaronpk did you notice? your "17 chars available " you calculated - and the "23 (vs 6 " in my post?
#
tantek
23-6 = 17
#
aaronpk
only after you pointed out the autolink change :)
#
tantek
well that's not how I expected to spend an hour + of my time before HWC SF
#
jjuran
what is PSL?
#
Loqi
A permashortlink (abbreviated PSL) is a URL using a short-domain that expands to a permalink; on the IndieWeb, PSLs use personal short domains to expand to the same person's personal domain, thus minimizing the fragility often associated with shortlinks https://indieweb.org/PSL
#
jjuran
what is PSC?
#
Loqi
A permashortcitation (or permashortid), abbreviated as PSC is a short non-hyperlinked citation to a post permalink, in contrast to a permashortlink which does hyperlink to a post https://indieweb.org/psc
#
tantek
pours one out for PSCs
#
tantek
I think Twitter just killed PSCs
#
aaronpk
thank goodness 😂
#
loqi.me
created /Web_of_trust (+232) "prompted by camb[m] and dfn added by camb[m]"
(view diff)
#
GWG
Coopertino
#
GWG
Excuse me
#
GWG
I was correcting pronounciation in another room
#
GWG
Hit the wrong window
calumryan joined the channel
tbbrown joined the channel
#
GWG
Correcting pronounciation is one of my things
tantek joined the channel
#
loqi.me
edited /iPad (+46) "tantek added "http://peopletakingpicswithipad.tumblr.com/" to "See Also""
(view diff)
#
KartikPrabhu
pronunciation*
#
gRegorLove
Correcting spelling is one of KartikPrabhu's things ;)
hi joined the channel
#
gRegorLove
aaronpk, Do you have stats on the # of tells Loqi hasn't delivered yet? That "hi" one got me curious.
#
GWG
KartikPrabhu: I realized I did that
#
aaronpk
ha yeah I can dig that up
#
aaronpk
604 undelivered across all servers and channels
[snarfed], tantek and j12t joined the channel
#
@t
Homebrew Website Club SF started with @Johannes_Ernst & quiet writing. 1 Million Webmentions (https://snarfed.org/1-million-webmentions) #IndieWeb party started last night in Baltimore, London, now we here! Come by @MozSF! We have cupcakes:) https://www.godaddy.com/garage/indieweb-facebook-opportunities/ http://tantek.com/t4sa3
(twitter.com/_/status/956348455244279809)
#
tantek
a-ha! new Twitter link-preview behavior detail
#
tantek
Twitter shows the the link-preview for the last link in a tweet **that has Twitter Card (or compatible) markup**
#
tantek
so because my permalinks DO NOT have Twitter Card metacrap, my PSLs are ignored by Twitter for link-preview purposes!
#
@t
Homebrew Website Club SF started with @Johannes_Ernst & quiet writing. 1 Million Webmentions (https://snarfed.org/1-million-webmentions) #IndieWeb party started last night in Baltimore, London, now we here! Come by @MozSF! We have cupcakes:) https://www.godaddy.com/garage/indieweb-facebook-opportunities/ http://tantek.com/t4sa3
(twitter.com/_/status/956348455244279809)
#
tantek
and it shows the link-preview for the 2nd to last link :)
#
tantek
aaronpk, you may be interested in ^^^
j12t and [jgarber] joined the channel
#
Loqi
[jgarber623] #80 RuboCop Cleanup Pt. 1
#
aaronpk
is that new? What did it do before?
j12t, eli_oat, bdesham_, tantek, renem and wagle joined the channel
#
tantek.com
edited /timeline (-5) "/* 2013 */ -beta"
(view diff)
tantek joined the channel
tantek, j12t and [miklb] joined the channel
#
tantek.com
edited /generations (+363) "/* Generation 2 */ general thinking in the room is that Fred Wilson is indieweb, and gen 2"
(view diff)
j12t joined the channel
#
tantek.com
edited /this-week-in-the-indieweb (+113) "Browse Recent Digests"
(view diff)
#
tantek.com
edited /Posts_about_the_IndieWeb (+40) "heading for the all the top links which otherwise people skim right over"
(view diff)
j12t and tantek joined the channel
#
staceydepolo.com
edited /GoDaddy (+344) "added sections and updates of current progress"
(view diff)
#
j12t
Looks like something's wrong with https://indiewebify.me/ "Web Sign In'. sdepolo failed to get it to pass for her site, although she links to and from Twitter. And it does not validate for me either, claiming that I'm not linking back from Twitter, but I do: twitter.com/Johannes_Ernst points to https://t.co/kcd30NSNtT which redirects to https://upon2020.com/ which redirects to https://upon2020.com/blog/ which has my rel=me
#
j12t
How can we debug this?
#
KartikPrabhu
did Twitter remove rel-me?
#
KartikPrabhu
nope, it still there
#
j12t
Actually it may be an indiewebify.me problem: I just killed all cookies on indieweb.org and managed to log in through Twitter without problems.
tantek, chrisaldrich and loicm joined the channel
#
tantek.com
edited /Facebook (+54) "/* How to wean yourself from */ moderate friending activity"
(view diff)
cweiske joined the channel
jeremycherfas joined the channel
#
@t
#ActivityPub (AP) is also now a @W3C Recommendation! https://www.w3.org/TR/2018/REC-activitypub-20180123/ Congrats @dustyweb @Tsyesika! Supported by @MastodonProject + interop like #indieweb site <—(#Webmention)—> http://fed.brid.gy <—(AP)—> Mastodon #openweb ... http://tantek.com/t4sa6
(twitter.com/_/status/956436348361158658)
tantek joined the channel
#
vanderven.se martijn
edited /ActivityPub (+1) "Proposed Recommendation → Recommendation 🎉"
(view diff)
EmreSokullu, loicm and tantek joined the channel
EmreSokullu, loicm and j4y_funabashi joined the channel
#
j4y_funabashi
what is ssl
#
Loqi
j4y_funabashi: Zegnat left you a message on 2017-11-29 at 5:53pm UTC: I was thinking about multiple people posting to a micropub endpoint not too long ago - https://indieweb.org/Micropub-brainstorming#Users_using_their_own_identities_to_log-in_to_a_community_blog.
#
Loqi
HTTPS is an abbreviation for Hypertext Transfer Protocol Secure, a form of HTTP where the communication is encrypted https://indieweb.org/ssl
marcthiele joined the channel
#
petermolnar
ssl is not https
#
petermolnar
this shouldn't be redirected like that
John_Ivan joined the channel
#
vanderven.se martijn
edited /subdomain (+271) "Used to have a subdomain, now canonically a path. Other family members still on subdomains."
(view diff)
marcthiele joined the channel
#
@rmondello
iOS 11.3 and macOS 10.13.4 include Service Workers — a powerful specification that allows background scripts to power offline web applications. iOS 11.3 also consults Web App Manifest when adding web apps to the home screen.
(twitter.com/_/status/956256845311590400)
[miklb] joined the channel
#
[miklb]
yes, exciting
[jgarber] joined the channel
[tantek] joined the channel
#
aaronparecki.com
edited /videos_about_the_indieweb (+120) "add stacey's video from HWC"
(view diff)
leg, eli_oat and [miklb] joined the channel
#
loqi.me
edited /Jekyll (+42) "[miklb] added "https://github.com/dumaurier/pwa_jekyll" to "See Also""
(view diff)
camb, snarfed and John_Ivan joined the channel
#
petermolnar
question about GDPR: the plain simple apache log format has the IP address as well - will that be a problem regarding the eu regulations?
#
sknebel
petermolnar: if you keep logs long-term, probably
#
petermolnar
define long term
#
sknebel
if I remember right current german court rulings (which aren't based on GDPR, but similar principles) say a week or two
#
petermolnar
I was expecting x years as long term
#
sknebel
e.g. uberspace doesn't give their customers visitor IPs at all
#
petermolnar
I can't even run fail2ban efficiently without a few months of logs including ips
#
sknebel
does your fail2ban really require to catch repeats that long-term?
#
petermolnar
it's set up as incremental ban
#
petermolnar
so first it bans for a few hours, then for a few days, depending on the last few hours, then for weeks, depending on days...
#
sknebel
ok, but that does not require full logs
#
sknebel
keeping a list of bad IPs for defense purporses is different from keeping logs of all your visitors
#
Zegnat
I find the IP adress-as-personal-information thing obvious on one hand, annoying on the other. Not sure how this would be for e.g. ban lists.
#
sknebel
(It's obviously also one of those "who's gonna sue you for it if you aren't a government or big company, if they have no idea if the court is going to agree this time" things)
#
petermolnar
the ban list is exactly what we're talking about
#
sknebel
I'd expect a ban list to be treated differently, since a) you as an operator have a clear reason why you need to keep it, b) normal users should not appear on it without reason, c) you collect very little auxillary data.
#
petermolnar
ok, but if you don't have the auxillary data, you can't tell why the ip is banned
[tantek] joined the channel
#
petermolnar
therefore they can request removal
#
[tantek]
Wow I had never heard the expression “cookie licking” before: https://medium.com/@steve.yegge/google-doesnt-necessarily-need-innovation-95cea96d0eeb
#
sknebel
I don't think keeping "banned because wordpress scanner" or "rate limit violation" would make a difference
#
[tantek]
That post and this one: https://medium.com/@steve.yegge/why-i-left-google-to-join-grab-86dfffc0be84 are full of gems explaining why and how bigcos (including silos) are slowing down in terms of innovation and doomed to eventually stagnate.
#
[tantek]
This is how and why we the indieweb can and will catch up and surpass the silos in terms of UX and innovation
#
sknebel
from the second one I found "Google has become 100% competitor-focused rather than customer focused." a good point
#
[tantek]
1. Our innovation is fundamental distributed and decentralized. Despite that yes we have a community, it’s all opt-in by the minute, everyone chooses to be here or not.
#
[tantek]
Nearly zero cost to staying/leaving/returning so you do you and innovate as it works best for your site.
#
[tantek]
2. We are constantly adding new innovators. New devs are showing up pretty much every week with their own perspectives and priorities. And we encourage everyone to document their own priorities publicly with Itches lists etc per /wikify
#
[tantek]
3. We focus any “centralization” work (read: standards) on the minimum necessary to make our actual innovations (UX, user features) interoperate.
#
[tantek]
This gets to something I’ve said several times when people want a new microformat for x or a new eel value for y - what is the usecase?
#
[tantek]
But really, what is the *interop* usecase? Without an interop usecase, no need to standardize.
#
[tantek]
new *rel value [lol]
#
[tantek]
“Interop usecase” is a broader way of saying what is both the publishing use case and consuming code usecase - you need both. Interop is why.
#
[tantek]
All of the above is also a good reason and driver to minimize feature sets of standards themselves. To prefer smaller standards. Minimum viable building blocks. Each of which provides utility on its own, and hopefully when combined provides new utility that by themselves they did not.
#
[tantek]
Lastly, by keeping the innovation as distributed as possible, it makes it less and less possible (read: more expensive, unworldly, and too weird looking to convince checkwriters) to “acquihire” or “acquishutdown” the indieweb. Another advantage over monocultures.
#
[tantek]
*unwieldly
snarfed joined the channel
#
[tantek]
Which is also why we must keep growing, mentoring, and embracing having more and more different thinking innovators. Makes it much harder to acquihire individuals as a way of shutting down a community (no evidence this is happening yet, unlike what happened to AS etc in mid/late 2000s with the old “social web” in crowd being hired away by Google, FB, Twitter etc)
#
[tantek]
Also a very good reason to avoid and discourage any kind of in-crowd vs not divisions, by doing the opposite as we do. Welcoming new folks “in” immediately and respecting and even empowering new/different opinions (even if imperfectly so, self included) as sources of breadth & strength for the community.
#
[tantek]
The flip side of that is that yes, just like the meta seeming contradiction but consistent in practice of not tolerating intolerance (per code-of-conduct), not tolerating monoculture attitudes, sales pitches, appeals to tradition, or worse appeals to “if you dont do it my old way then why do you hate me”.
snarfed1 joined the channel
#
[tantek]
The answer is: false dichotomy.
#
[tantek]
Yes any of us can innovate on our own sites differently than anyone else, and no it does not mean we hate nor does it take away from other approaches. It means we are innovating with our UX and user features in a way the makes the most sense and importance to ourselves (self dog food). That exploration of diversity by implementing and using (show rather than tell), strengthens the community and frankly makes it increasingly more fun and intere
#
schmarty
^ tantek: this seems very on-topic and probably would make a very good post!
#
schmarty
(i mean for main rather than just -dev)
#
schmarty
tantek++
#
Loqi
tantek has 23 karma in this channel (419 overall)
EmreSokullu joined the channel
#
aaronpk
tantek++ I like it
#
Loqi
tantek has 24 karma in this channel (420 overall)
#
aaronpk
Oh boy [eddie] I've got a lot of stuff to review from you haha
#
Loqi
aaronpk: lol
#
cleverdevil.io
edited /reader (+928) "/* IndieWeb Examples */"
(view diff)
[cleverdevil] joined the channel
#
[cleverdevil]
I'm reallllly enjoying it 🙂
#
[cleverdevil]
More pretty screenshots of Together.
snarfed joined the channel
#
voxpelli
Together looks really neat, nice to see good progress in that space since I last was around :)
#
Loqi
voxpelli: tantek left you a message on 2017-09-26 at 9:48pm UTC: I'm experimenting with a "follow" webaction on my homepage - take a look and let me know if you think /indie-config needs any changes to support handling "follow" and passing it to your website to handle
#
KartikPrabhu
[cleverdevil]: seems to be borrowing Google's aesthetic
#
KartikPrabhu
but looks great
[eddie] joined the channel
#
[eddie]
aaronpk: just keeping you on your toes 😉
#
GWG
I think aaronpk is on his toes enough
#
Loqi
[aaronpk] you're sending the webmention with a source URL that contains the comment ID, which is great but almost nobody else does that
#
GWG
What can we do with that greatness?
#
aaronpk
At the very least it can tell me that my comment has been published
#
GWG
It doesn't?
#
GWG
Remember that WordPress has moderation.
#
GWG
We don't have a way to indicate that, do we?
chrisaldrich joined the channel
#
GWG
I wonder about that the other day. If a comment is created, but held for moderation, how should a Webmention endpoint respond?
#
GWG
Maybe a 202?
#
Zegnat
Webmention endpoint only communicates whether the mention itself was accepted, not what happens after accepting.
#
GWG
Does moderation constitute an asynchronous response?
#
Zegnat
But you could display more information on the status page.
#
GWG
We don't have a status page
#
Zegnat
200, 201, and 202 all have specific meanings: https://www.w3.org/TR/webmention/#receiving-webmentions
#
[cleverdevil]
@KartikPrabhu yeah, Together uses a Google Material Design style/theme.
#
KartikPrabhu
[cleverdevil]: yeah it was a minor point. The "circle with the +" always reminds me of Google products. But I am sure that can be changed once Together is more mature
[tantek] joined the channel
#
[cleverdevil]
Yup. Still a lot of work to do.
AngeloGladding joined the channel
#
snarfed
ugh, huffduff-video is serving >3T/month. guess it's time to go back and crunch logs again
snarfed and raretrack joined the channel
#
j4y.co
edited /User:J4y_funabashi (+2) "/* major */"
(view diff)
#
@_am1t
Experimentation continues with IndieWeb projects – over to Webmentions. It is a really wonderful concept which enables responses to a post to be written on one’s own websit. As Jeremy Keith wrote in one of his posts: Basically, it’s an equivalent to pingback. Let’s say I wr…
(twitter.com/_/status/956619521279844352)
EmreSokullu and [cleverdevil] joined the channel
#
[cleverdevil]
So, debugging an issue with manton on his new Sunlit 2.0 iOS app.
#
[cleverdevil]
When his app attempts to publish to my Known instance via Micropub, for some reason the HTML is getting escaped rather than accepted as the raw content.
#
[cleverdevil]
I've got a simple test script that publishes HTML via Micropub and I am seeing the same thing.
#
[cleverdevil]
Not sure if its a bug in Known, or if we're *both* doing something silly.
#
sknebel
[cleverdevil]: try even more manual, send a request with curl? or share the script and we can take a look
#
[cleverdevil]
Its super simple.
#
[cleverdevil]
I know its worked in the past, so I am guessing its a regression in Known.
#
sknebel
[cleverdevil]: you can't create HTML posts with formencoded requests
#
[cleverdevil]
Ah, well that explains that 🙂
#
sknebel
so known is correct here, creating a plaintext post
#
[cleverdevil]
I'll convert my script to use JSON and see if that works.
#
[cleverdevil]
And I will report back to Manton.
snarfed, j4y_funabashi, KartikPrabhu, chrisaldrich, j12t and j12t_ joined the channel
#
www.boffosocko.com
edited /Webmention (+2132) "moved publishing software and services from wm-developer page (reduced duplicates); fixed TOC table depth for publishing software section"
(view diff)
#
www.boffosocko.com
edited /Webmention-developer () "(-3178) moved publishing software and services to main wm page (and reduced the duplicates); tweaked TOC level for Libraries section"
(view diff)
[manton] joined the channel
#
[manton]
This Micropub plain text vs. HTML issue is kind of similar to a discussion we had here about supporting or requiring Markdown. In other words, should a server be able to limit the formats it accepts.
#
Loqi
[manton]: gRegorLove left you a message 2 weeks, 1 day ago: Where can we file bugs for micro.blog?
#
Loqi
[manton]: gRegorLove left you a message 2 weeks, 1 day ago: https://gregorlove.com/2018/01/i-signed-up-for-a/
#
[manton]
I don't really see anything in the spec that says HTML isn't allowed in form-encoded POSTs.
#
KartikPrabhu
[manton]: sure. but the server is then free to interpret them as plain-text
#
KartikPrabhu
in a form-encoded POST request there is no way for the server to distinguish between plain text and markup
#
[manton]
That's true. I'm just trying to understand what is in the spec vs. what is an implementation detail, like Known not accepting HTML.
#
sknebel
it's indirectly in the spec by saying that it follows mf2 semantics, where unescaped html only happen in that nested form
#
sknebel
so all other text is assumed to be plain text
#
sknebel
(you are right though, the spec could call that out clearer. guess that was "too obvious" at some point?)
#
[manton]
It's not clear to me. I think having 2 forms complicates things a little... As one example, the spec says a Micropub client that creates posts _must_ support sending form requests. But what if a client has to send HTML?
#
[manton]
If I switch to sending JSON, then technically my client won't be conforming anymore.
[cleverdevil] joined the channel
#
[cleverdevil]
It would be nice to be able to send HTML via the form-encoded method.
#
sknebel
you'd have to ask one of the people involved with the spec-making process about the "conforming" bit (e.g. aaronpk)
#
sknebel
the form-encoded form is the simpler one that from what I understand intentionally doesn't support all cases since it got to complicated otherwise
#
[manton]
Yeah, I love the simplicity of it.
#
[manton]
Maybe if there was a way to query the server for whether it preferred or supported form-encoded vs. JSON, that would help. Not sure... We don't need to debate it now, necessarily, but I do want my apps to be more compatible.
#
sknebel
I think the most common behaviour for newly written clients that try to be compatible is to use form-encoded where possible and switch to JSON when it's not
KartikPrabhu joined the channel
#
sknebel
on the other hand, I have no overview how large the set of servers that don't support JSON is, I *think* the ones that are not just private projects largely do JSON
#
www.svenknebel.de
edited /Micropub/Servers (+1) "/* Implementation status */ wordpress issue for JSON is closed, so checkmark that"
(view diff)
tantek joined the channel
#
sknebel
!tell GWG could you review the wordpress column in https://indieweb.org/Micropub/Servers#Implementation_status and bring it up to date if it is not? Thanks!
#
Loqi
Ok, I'll tell them that when I see them next
#
[cleverdevil]
So, I modified my script to use the JSON format and it works now, validated.
#
aaronpk
It'll get really messy if we try to make the formencoded content anything except plain text.
#
aaronpk
You *could* have a server that interprets the content as markdown, since markdown is close to plain text, but only if it returns the markdown source when queried, and only if you're okay with some clients showing the markdown source when editing
#
aaronpk
If you look at my text notes with bullet lists, I actually author those by typing * but my server renders them with •
#
aaronpk
The form encoded request is meant to be the simple version just to get started, but as soon as you want to do any "fancy" stuff like posting checkins or dealing with html content then JSON lets us be more explicit about the request
#
[cleverdevil]
Fair enough. Its definitely still simple. Just not *as* simple.
#
[cleverdevil]
One piece of feedback on the spec: it could use a lot more fully-defined examples with MF2 JSON.
#
[cleverdevil]
In fact, if there could be an example that exhibits every single property for every single type of content that can be created, that would be amazing.
#
aaronpk
There might be more in the wiki right now
#
aaronpk
if not, that's the place to add them
#
aaronpk
At least for now
#
[cleverdevil]
Makes sense.
snarfed, sebsel and [eddie] joined the channel
#
[eddie]
!tell aaronpk I added micro.blog’s JSON timeline feed to Monocle. It’s the first feed I’ve noticed the character encoding issue with. Is this an issue in Indigenous, Monocle or Micro.blog? https://indiewebcamp.slack.com/files/U433EQWG2/F8YQM7JSF/image_uploaded_from_ios.jpg
#
Loqi
Ok, I'll tell them that when I see them next
#
[eddie]
eddie mentioned a file: Image uploaded from iOS.jpg.
[manton] joined the channel
#
[manton]
eddie Let me know if there's anything I need to look at. Looks like the feed is correct on first glance.
#
[eddie]
Thanks manton. I’m guessing it’s just deciding whether the decoding happens in a Microsub Server or a reader interface. I think, as you said, the feed is probably fine
#
aaronpk
[eddie]: oh that's delightful
#
Loqi
aaronpk: [eddie] left you a message 22 minutes ago: I added micro.blog’s JSON timeline feed to Monocle. It’s the first feed I’ve noticed the character encoding issue with. Is this an issue in Indigenous, Monocle or Micro.blog? https://indiewebcamp.slack.com/files/U433EQWG2/F8YQM7JSF/image_uploaded_from_ios.jpg
#
aaronpk
at least my version
#
aaronpk
check the `entries` table on yours for unique="@293775"
#
[eddie]
Interesting. My content/text has the encoded characters
#
[eddie]
But the content/html one is fine
#
[eddie]
I wonder if it has something to do with the way my PHP is configured
#
[eddie]
Well, not that big of a deal for now then 🙂 Probably something wrong with my local monocle
#
aaronpk
which micro.blog URL did you add?