#dev 2021-02-02

2021-02-02 UTC
#
KartikPrabhu
really doesn't matter if HTML is labelled as "programming" or not. It does what it is supposed to do
#
KartikPrabhu
a bicycle is not a jet engine but a bicycle is still useful
#
KartikPrabhu
so who cares
#
aaronpk
especially because "programming" actually has many different definitions
#
btrem
I suppose a bicycle repair shop might care.
#
KartikPrabhu
someone should go to a bicycle shop and yell "this is not a jet engine"
#
KartikPrabhu
I bet they would just be thrown out :P
#
btrem
Of course, I meant they'd care if you brought in a jet engine and said, "fix this please".
#
KartikPrabhu
exactly! they would tell you to go somewhere else
#
btrem
I get that person would get thrown out, too.
#
btrem
So they both get thrown out!
#
KartikPrabhu
similarly, if you want a turing complete language for some use then don't use HTML
#
KartikPrabhu
if you want to markup a document use HTML
#
btrem
I don't think anyone argues with that.
#
aaronpk
apparently people do
#
KartikPrabhu
completely useless "argument"
#
btrem
? That you should use HTML to mark up a document?
#
aaronpk
(*ahem* React)
#
aaronpk
or more generally, SPAs that aren't actually applications, they are documents
#
btrem
What is an SPA? (No, Loqi, I'm not asking you!)
#
Loqi
A single-page application (SPA) is a web application or web site that fits on a single web page with the goal of providing a user experience similar to that of a desktop application https://indieweb.org/SPA
#
btrem
Oh, but Loqi knew!
#
btrem
Well hush my mouth.
[fluffy] joined the channel
#
[fluffy]
Hmm, so, in my h-card I purposefully overly URL-encode my email address to e.g. “mailto:fluffy%20beesbuzz%2ebiz” because it’s an easy way to defeat noncompliant things like spam harvesters from getting my address, but allows worthwhile user-agents like web browsers to work right. But I’m noticing that having that causes my webmentions to fail validation on some endpoints, such as Barry Frost’s site. Does anyone have any opinions on be
#
[fluffy]
options?
#
[fluffy]
like his webmention validator tries parsing the mailto: link but doesn’t urldecode it.
#
[fluffy]
And then it ends up rejecting the entire webmention.
#
[tantek]
sounds like a good test case for webmention.rocks!
#
[tantek]
fluffy, that URL encoding seems fine to me. it shouldn't cause an HTML parser to fail
#
[fluffy]
yeah, that’s waht I thought. Maybe I’ll reach out to Barry to suggest he urldecode things. 🙂
#
[fluffy]
or to at least not reject an entire h-card based on a single attribute failure.
#
aaronpk
seems weird that it would fail on the whole thing for a property that isn't usually used in the display of webmentions
#
[tantek]
yup, worth suggesting a test case for receivers: https://github.com/aaronpk/webmention.rocks/issues
#
[fluffy]
oh wait! the error came from fed.brid.gy
#
[fluffy]
not from barry’s endpoint
#
[fluffy]
wait no, I misread my misreading
#
[fluffy]
There were two errors adjacent to each other, one from bridgy fed (that always occurs on my reply posts for some reason) and then Barry’s endpoint, and my error log doesn’t always make it that easy to read. 🙂
#
Loqi
[fluffy-critter] #41 h-card validation tests
#
[fluffy]
feels vaguely funny how Slack’s github integration always displays the avatar of the repo owner, rather than the issue opener.
#
vilhalmer
I think that just comes from github's generated twitter card metadata
#
vilhalmer
and/or opengraph
#
aaronpk
right, github should be putting the issue creator's info there
#
aaronpk
not the repo creator
#
vilhalmer
if only there were some sort of open standard for embedding machine parsable information into html instead of each inventing their own
KartikPrabhu joined the channel
#
jacky
if only
#
[KevinMarks]
I like Lea's response to HTML isn't programming - mavo.io
#
[fluffy]
That feels… angular-ish
btrem joined the channel
#
jacky
very random but [manton] is there a way to remotely sign up for micro.blog?
#
jacky
I'm thinking of 'advanced' cases of a social reader
#
jacky
and one that kinda comes to mind is the ability to do something like a 'introspective' view
#
jacky
like if you provide it a URL, it'd render said post under the reader and attempt to do some more stuff like subscribe briefly to the websub hub shown to potentially give information about incoming posts
#
jacky
or provide a simpler interface to engage with it (similar to how mastodon kinda does it)
#
aaronpk
interesting
#
jacky
I don't want to bake some of this stuff into my site - I'd rather have it on a standalone tool so I can then 'link' to it and have people use it on their sites as well
#
jacky
I think the biggest thing is that 'engagement' (via responses, not like metrics) would only be surfaceable if they provide it on a page (like there's no means of querying replies for a particular URL)
#
jacky
(that's kinda where having a Webmention endpoint be queryable would come into play imo)
#
jacky
I'm thinking about this stuff b/c I want it mainly on my site but also to make it clear that individual sites _can_ do this and in a simpler way
#
aaronpk
you can enter a feed or post URL here to see what it looks like in monocle https://monocle.p3k.io/preview
#
aaronpk
what are you thinking beyond that?
#
jacky
I should probably remove that h-feed bit at the bottom lol
#
jacky
but maybe a spot to provide a means for replying (which would prompt Web sign-in, if not already) and perhaps reactions (if any)
#
jacky
something to bring it to a similar view to what a singular tweet would look like tbh
#
jacky
tbh I don't know if that's worth it yet lol
#
aaronpk
there's definitely some fun implications there
#
aaronpk
replying by posting to your own site via micropub? replying by having this tool create a url and send a webmention itself?
#
jacky
the latter would be something that could people who don't even have micropub fully set up
#
jacky
"send-me-a-webmention-to-this-page" as a service
#
jacky
going to think about this more on the bike
ShadowKyogre and [jacky] joined the channel
ShadowKyogre and [snarfed] joined the channel; ShadowKyogre left the channel
#
[snarfed]
hey aaronpk, low priority, when you get a chance, could you check if wm.io’s client HTTP lib supports SNI now? it didn’t back in 2014, which was (one of) the reasons bridgy uses brid-gy.appspot source URLs. https://github.com/aaronpk/webmention.io/issues/14
#
Loqi
[snarfed] #14 not seeing some brid.gy source URLs
#
[snarfed]
evidently older version(s) of openssl you were using back then didn’t support it: https://chat.indieweb.org/2014-01-04#t1388877077000000
#
Loqi
[aaronpk] looks like system openssl version is 1.0.0e on the server and 1.0.1e on my mac
#
aaronpk
Oh huh
[xavierroy] and [benatwork] joined the channel
#
[benatwork]
Submitted a tiny change to the PHP mentions client https://github.com/indieweb/mention-client-php/pull/39
#
[benatwork]
PHP merrily reversed the parameter order of a function because ¯\_(ツ)_/¯
#
Loqi
[benwerd] #39 Fixed bug with implode parameter order
#
@maznu
↩️ It's a subject that keeps coming up time and again on the Orange website… https://utteranc.es/ — comments as github issues https://webmention.io/ — standards track? https://github.com/tessalt/echo-chamber-js — "lol" https://lisakov.com/projects/open-source-comments/ — comparison table of other options
(twitter.com/_/status/1356487185113964548)
#
@mdrockwell
Installed the Webmentions plugin for WordPress. Really digging it, even with just a little bit of experience with i… https://twitter.com/i/web/status/1356385122891796480
(twitter.com/_/status/1356385122891796480)
deathrow1 joined the channel
KartikPrabhu, schmudde, marinin[t], [xavierroy] and [barryf] joined the channel
#
[barryf]
[fluffy] Sorry for the failed webmention. Your h-card is perfectly valid of course. The exception comes from the indieweb-authorship Ruby gem which my server uses. I’ll fork and see if I can add some exception handling.
#
[barryf]
I’m not sure if the author, deeden, drops in here?
[srushe] joined the channel
#
[srushe]
I do. I generally lurk.
#
[barryf]
Hi Stephen. My server didn’t parse Fluffy’s webmention because of `URI::InvalidComponentError: unrecognised opaque part for mailtoURL: fluffy%40beesbuzz%2ebiz` in Indieweb::Authorship.identify. I’ll see if I can recreate locally.
#
[barryf]
If I can fix I’ll send you a patch a bit later on.
lanodan joined the channel
#
[srushe]
I’ve got it recreated locally. I’m not sure what to do here. The microformats gem is raising the error. I can rescue it fairly easily but all I could really do them is return an empty response. If you find a better solution I’ll be happy to hear it 😄
#
@GR36
Exploring webmentions on my blog. Works most of the time, but is a little spotty in places. http://Micro.blog seems to only work some of the time, and replies don’t show up as my http://micro.blog account. But I’m not sure they are supposed to? @help
(twitter.com/_/status/1356557917470744577)
[KevinMarks] and schmudde joined the channel
#
@FlorianUhde
Spend the last days working on connecting to the #IndieWeb and webmentions at https://florianuhde.com Small babysteps, but fun and with plenty of potential
(twitter.com/_/status/1356583482772889600)
shoesNsocks, shoesNsocks1, joaweyang, [KevinMarks], [schmarty], [tantek], [timothy_chambe], [snarfed] and [fluffy] joined the channel; ShadowKyogre left the channel
#
[fluffy]
[barryf] oh hi, didn’t realize you were on this Slack 🙂
#
[fluffy]
well, chat in general (but you happen to be on it via Slack)
#
[fluffy]
[srushe] Sounds like it’s turtles all the way down. Is there a particular exception that could be caught? Or maybe the mf2 parser author(s) need to be brought in too 🙂
#
[fluffy]
like ideally the card validation wouldn’t fail because of a single “malformed” optional attribute, but also the attribute isn’t malformed.
#
[fluffy]
(and many people do purposefully malform that particular attribute!)
[srushe] joined the channel
#
[srushe]
[fluffy] I can work around the issue if I explicitly decode the html before I send it to the Microformats parser, but that “feels” wrong to me. I think it should probably be up to the Microformats gem to handle the encoded html, but that may just me passing the buck. I’ll look to put up a version that works around the issue for now
#
[fluffy]
Whatever is parsing the HTML should also be decoding the URLs. And that does sound like it’s the Microformats parser’s responsibility.
#
[fluffy]
although that also heads into a gray area, like when handling HTTP URLs, should you round-trip it through a urldecode/urlencode phase or whatever? That leads to fun corner cases.
#
[fluffy]
like then you can’t support path-based things with /s in a path component, like `http://foo.example/blog/category/books%2fmovies/`, and I’ve seen much disagreement among web frameworks as to whether that should be supported in the first place (for example, wsgi explicitly disallows it).
#
[fluffy]
(or rather the wsgi spec says you must url-decode the incoming URL components before passing them along to the application)
#
[srushe]
Exactly. My first stab at this this morning is a simple fix, but I’m certain that there will be weird cases where it’s a bad idea. I may pull the gem source down tonight and see if I can solve the problem there.
#
[fluffy]
%-encoding is one of those things where things were casual enough in the beginning that formalizing things now is difficult to do right. 😕
#
[fluffy]
anyway, time for day job meeting, ugh
marinin[t] joined the channel
#
jacky
lol, I might end up doing a little 'blog storm' of the changes I have to make to my site for https://aaronparecki.com/2020/12/03/1/indieauth-2020
#
jacky
probably won't be today (so much to do!)
#
Loqi
[Aaron Parecki] IndieAuth Spec Updates 2020
#
[tantek]
Curious if folks have thoughts about jasnell's thread: https://twitter.com/jasnell/status/1356460552768212992
#
@jasnell
All of the necessary pieces are still available. It's still possible to build these types of systems without much difficulty. It's not a technical problem, it's a logistical one. The difficulty lies in attracting the users and guarding/policing the spaces so that they remain safe
(twitter.com/_/status/1356460552768212992)
#
@jasnell
... you see, developing the technologies is the easy part. Bootstrapping the community that uses them is hard. And FB and Twitter were investing heavily in owning the space with their own walled gardens. FB especially. We knew the standards didn't mean squat if they wouldnt adopt
(twitter.com/_/status/1356460551061131264)
#
jacky
not wrong there at all imo
#
jacky
Like we can have the most ideal solution
#
jacky
but if only tumbleweeds are the users, it's useless (which is sad)
#
jacky
we slightly get through this because we bridge into any community we have an interest in (or if it's easy to)
#
[tantek]
I disagree with "developing the technologies is the easy part". It's only easy if you're ok with complex solutions. Simple(r) technologies are ironically harder to develop.
#
jacky
that depends when you come in the process, no?
#
[tantek]
I disagree with "Bootstrapping", it's a regressive libertarian metaphor
[benatwork] joined the channel
#
[tantek]
I agree that growing a community is hard, which is one of the reasons aaronpk and I made a deliberate decision to start/grow *this* community, instead of the 2010 FSWS
#
[benatwork]
“Starting a business”
#
[benatwork]
Or just “starting / growing” in the non business sphere
#
[schmarty]
i argue with the framing of "developing the technologies" as if there is one correct set of technologies and the problem is that the world won't pick them up and use them.
#
[tantek]
and I disagree strongly with "We knew the standards didn't mean squat if they wouldnt adopt", actually quite the opposite. So much early time on "open social web" standards was *wasted* trying to cater to BigCos who blew it all off anyway
#
[benatwork]
tantek++ basically, we need to work *despite* them.
#
Loqi
tantek has 17 karma in this channel over the last year (83 in all channels)
#
[schmarty]
there's a case to be made that "developing _a_ technology" is easy because, sure, you pick some constraints and you solve for X.
#
[tantek]
I also disagree with "All of the necessary pieces are still available", because no, you don't and can't know that until you have actually built the system that solves user's problems
#
[tantek]
also disagree with "It's still possible to build these types of systems without much difficulty." — are you freaking kidding me? "without much difficulty" lololol
#
[schmarty]
my takeaway from this thread is "no one should build technology because we once tried to solve the wrong problems and it didn't fix anything."
#
[tantek]
this is also wrong: "It's not a technical problem, it's a logistical one." (yeah blame ops, nice try)
#
[tantek]
if a technology doesn't operationalize well, it's the fault of the technology design
#
[schmarty]
i read "logistical" in that usage as being about getting people to adopt the new tech.
#
[tantek]
wowow as if people are "logistics" yikes that's even worse of a capitalist framing 😞
#
[schmarty]
yeah the next sentence clarifies "logistical problem" as: "The difficulty lies in attracting the users and guarding/policing the spaces so that they remain safe"
#
[schmarty]
Growth Hacking x Moderation
#
[tantek]
which brings me to one of the worst wrong things in this thread. yes that one
#
[tantek]
like are you kidding me? What's wrong with "attract users" then "police them"? anyone?
#
[tantek]
this is what I'm saying, that's a horrible default authoritarian framing
KartikPrabhu joined the channel
#
[schmarty]
to borrow a phrase i read in here recently: it's like abducting the entire city into your party and getting rid of anyone who harshes the vibe.
#
jacky
hm I do agree with y'all
#
[schmarty]
i think this exact thing is playing out with ActivityPub. it's _a_ technology that enables (and encourages) certain kinds of experiences and excludes (or makes difficult) others. even so, there is some _frothy_ evangelism around it as being The Way that social networks should work from now on.
[timothy_chambe] joined the channel
#
aaronpk
"some"?
#
[schmarty]
as someone who has not observed _all_ of the evangelism, i can't speak to whether _all_ of it is frothy. 😂
#
aaronpk
i was more speaking to the quantity than the froth
#
jacky
lmfao
#
[schmarty]
fair! there is both a high quantity and high intensity of ActivityPub-for-your-problems evangelism.
#
[tantek]
that kind of "high quantity & intensity evangelism" has happened before, with other monolithic standards efforts.
#
[tantek]
on a positive note, I found out that WebWeWant.fyi accepts and handles webmentions in response to any of the "wants" posted there!
#
[tantek]
[KevinMarks] you might be interested in this in particular: https://webwewant.fyi/voting/
[KevinMarks] joined the channel
#
[KevinMarks]
I can see how jasnell end up there as he has been adding corner cases to AS for years, but AP then built half thought out apis on top of it and claimed that they were done.
#
aciccarello[m]
Oh cool, so a webmention is a vote 👍️
#
[tantek]
KevinMarks, adding corner cases by way of public iteration with a community of implementers is one thing, but that absolutely did not happen with AS
#
[tantek]
to be fair to AP, it was an iteration of what was implemented in pump.io so at least it had *some* grounding in real world public implementation
#
[tantek]
whereas AS never got off the ground in terms of an ecosystem of publishers & consumers
#
[KevinMarks]
Well, it did originally but then the community dissipated
#
[KevinMarks]
And Monica was a key part of implementing it in many places.
#
[tantek]
KevinMarks, never saw any functional consumers
#
[tantek]
just a bunch of different publishers
#
[tantek]
also no test suite
#
[tantek]
there was only a handful of publishers. there was zero actual *ecosystem*
#
[tantek]
especially when it made the jump to JSON
#
[tantek]
there were a few AS1/AtomXML consumers (I mean that's why I added the object-types to my Atom initially)
#
[tantek]
I think that was a side-effect of OStatus using AS1/AtomXML
#
[tantek]
(the existence of a few consuming apps/sites)
#
[tantek]
the lack of a test suite meant there was nothing to stop entropy from rotting the implementations that Monica or anyone else did
#
[tantek]
(at BigCos that is, where entropy destroys things in every iteration)
shoesNsocks joined the channel
#
[tantek]
lack of test suite -> lack of regression tests etc.
shoesNsocks, ShadowKyogre, miklb_ and btrem joined the channel
#
GWG
aaronpk: You said you'd considered adding the ability to show posts by city/location. Other than time, what stopped you?
#
GWG
Building the feature, trying to think about what I might miss
#
aaronpk
i haven't figured out how i want that kind of navigation to work
#
aaronpk
also i have a mess of location names that i'd need to sort out
#
aaronpk
depending on the source of the city name sometimes it's localized and sometimes it's in english
ShadowKyogre left the channel
#
GWG
aaronpk: I found that as well. I figured I'd use the WordPress term merge capabilities to manage that if needed
#
aaronpk
some day i'm going to figure out how to speed up my activitypub delivery so it doesn't take like 5 minutes
#
aaronpk
some of that is going to be being more proactive about marking servers as offline and not sending to them in the first place
#
[tantek]
"take like 5 minutes" would seem to indicate a scaling problem with that architecture
#
aaronpk
it sends them one by one right now
#
aaronpk
100% sequential
#
jacky
I was gonna ask if that was the case
[snarfed] joined the channel
#
[snarfed]
totally reasonable for a single site
#
aaronpk
the question is: is it worth trying to speed it up, and if so, what's the least effort way to do that
#
jacky
by pushing it to a separate service ;)
#
[snarfed]
uh he said least effort, not most 😂
#
[snarfed]
aaronpk one thought is to priority order the inboxes you’re delivering to. won’t improve total time, but if it hits the key ones in the first 30s or so, you might not care
#
aaronpk
true, also i need to prioritize other tasks my site does over these for sure, like if i send a reply while it's in the middle of pushing them out, i want the new reply webmention to go out first
alex11 joined the channel
#
aaronpk
i will say, not providing default HTML/CSS for showing webmentions from webmention.io has been working out pretty well cause people are doing cool things like this https://twitter.com/MaximeHeckel/status/1356737238080622594
#
[tantek]
very cool. more facepile examples