2017-03-14 UTC
[cleverdevil] joined the channel
# 00:02 tantek.com edited /chat (+414) "fix dfn to be more practically applicable / useful, indieweb examples none so far, silo examples stub, xkcd" (
view diff )
# 00:03 Loqi chat is informal messaging offered by numerous services, a few of which interoperate or bridge with each other, and also a set of brainstorms for what an amazing indieweb friendly chat web app/site could do https://indieweb.org/chat
# 00:18 tantek hey there was no need to delete the existing Glitch
# 00:18 tantek skenebel - only mistakes should be deleted. Glitch was an actual thing, and thus ok to move to another page (rename)
# 00:20 tantek we should only be deleting mistakes and redirects to make room for actual pages
# 00:23 sknebel I can create a page manually with the old content
# 00:34 tantek sknebel: also good to save any prior names when you rename things so old (textual) references are still findable
ben_thatmustbeme, bnvk, tantek, barpthewire, deathrow1, gRegorLove, cweiske, [kevinmarks] and miklb joined the channel
# 13:01 tantek bnvk - yeah I think we discussed it a while ago in IRC - try our IRC log search!
# 13:17 bnvk aaronpk: meaning you think it's kind of stagnant / not worth following?
# 13:19 aaronpk but the chances of major players like freenode switching to it are probably slim
# 13:19 aaronpk and there are also totally separate efforts going on like Matrix
# 13:20 bnvk ah, so that's your overall feeling about IRC these days... noted ;)
# 13:20 aaronpk matrix is kind of a complicated protocol, but they do seem to be doing a lot of things right
# 13:21 bnvk what do you think about their @username:domain.com syntax?
# 13:22 bnvk heh, me too. I made a strong impassioned argument to Matthew about that- but he wasn't having it :/
# 13:23 aaronpk I'm not sold on Matrix either, but I'm impressed they've gotten this far
# 13:24 aaronpk what i would really like to see is a protocol that is much easier to implement from scratch
# 13:24 aaronpk Like how with webmention you can implement it incrementally and still participate in conversations
singpolyma and tantek joined the channel
# 13:35 aaronpk I do keep thinking what an indieweb chat protocol would look like
# 13:36 Loqi chat is informal messaging offered by numerous services, a few of which interoperate or bridge with each other, and also a set of brainstorms for what an amazing indieweb friendly chat web app/site could do https://indieweb.org/chat
# 13:39 aaronpk the pieces are pretty well in place for that to work. The main difference between that and Matrix is that as written, the indieweb on relies on a single coordination server to distribute the messages
KevinMarks joined the channel
IRCFrEAK, tantek and KevinMarks joined the channel
leg and KevinMarks joined the channel
KevinMarks and tantek joined the channel
# 18:19 aaronpk !tell cweiske can I ask you a few questions about shpub when you get a chance?
# 18:19 Loqi Ok, I'll tell them that when I see them next
# 18:45 bear I so want to figure out how to get webmentions to be the backbone for a distributed chat system
# 18:46 tantek bear, just figure out a Webmention to XMPP bridge ;)
# 18:46 aaronpk bear: we can totally do it, and it would work great using an aggregator (like indienews) to coordinate the conversation
# 18:46 bear right - it's the coordination that gets me stuck
# 18:47 aaronpk i was reading the matrix spec about server-to-server federation, and they have a bunch of stuff about how conversations don't actually "live" in any one place
# 18:47 bear haha - you can't be in the XMPP world for as long as I have without having a good sense of humour
# 18:47 aaronpk and all the different players can backfill a channel from any other server
# 18:47 bear right - it's the deduping and discovery of new chat items that bogs me down
# 18:48 bear which feels like a solved problem with something like superfeedr and even xmpp muc
# 18:48 singpolyma webmention-to-XMPP would probably bridge to pubsub or MUC, in which case conversations *do* live in one place
# 18:49 bear and I know it's a solved problem with NNTP
# 18:49 bear singpolyma - agree, if they bridged to pubsub then any new entry could be federated to others by listening to that topic
# 18:50 bear hell, I know that back in the day of even FidoNet it was solved by a new node pulling in the starting list of topics to seed their cache
# 18:50 singpolyma probably the right model would be PSHB-to-PubSub (prosody has a module for that), get that working with a nice social networking client like Movim, then layer on webmention to that
# 18:50 bear which would be like a merger of NNTP and Torrent
# 18:51 singpolyma bear: Yeah. I have an older friend who is constantly saying "those who don't know their history are doomed to reinvent it badly over HTTP"
# 18:51 bear singpolyma - that is what I say almost weekly
# 18:52 tantek format corollary: those who don't know their history are doomed to reinvent it as JSON(LD).
# 18:52 Loqi tantek has 4 karma in this channel (321 overall)
# 18:52 tantek previously: those who don't know their history are doomed to reinvent it as XML
# 18:53 bear a topic would then become a dynamic article with each chat equal to a paragraph - then webmention and fragmention could be used
# 18:54 bear well, XML I would consider old enough that it could be what should have been used by many folks - but i'm not going to fall on that conversation-sword in this generation of developers any more
# 18:54 tantek previously previously (especially in gov/edu circles) those who don't know their history are doomed to reinvent it as SGML
# 18:55 tantek aaronpk, something that does a better job with i18n string support
# 18:55 tantek because all the name-map crap is a total hack
# 18:56 tantek and makes it so much more work that anything other than monodirection text is diminished
# 18:57 tantek which BTW shipped on by default in FF52 last week!
# 18:57 bear I think i'm going to start calling WebAssembly a trigger word and having my computer just reboot when it appears
# 18:58 tantek but bear, the devtools now have a disassembler! you gotta love that
# 18:58 singpolyma bear: just wait until we have a JVM compiled to webassembly so you can run stuff in a sandbox!
# 18:58 tantek like disassembly into a text format you can then *write* and assemble, if you were missing writing assembly by hand ;)
# 18:58 bear so I can have my XML loving JVM stuff all inside of a browser
# 18:59 bear haha - that's it - i'm going to start a bespoke hand assembly shop
# 18:59 Loqi bear has 14 karma in this channel (181 overall)
# 18:59 bear I think I still have some books from the 80's in my basement that I can mine for months
# 19:00 tantek bear, that's not a bad business model, given how much really slow JS there is out there, and site will be looking for JSEO
# 19:00 bear yea, I was mostly joking - but in reality if will soon be just like nodejs --> reactjs path
KevinMarks and cweiske joined the channel
# 19:19 Loqi cweiske: aaronpk left you a message 1 hour ago: can I ask you a few questions about shpub when you get a chance?
# 19:21 aaronpk you saw my comment on github, we can keep that conversation thread there
# 19:23 aaronpk now that I've finished the client test suite, do you think you could run through the tests to verify your implementation report?
# 19:24 aaronpk (the W3C group voted today to request the Micropub PR transition only after two or more reports are verified against the test suite)
barpthewire joined the channel
# 19:50 martymcguire[m] aaronpk: just submitted an updated implementation report for screech from running through micropub.rocks client tests. should i include a link in there to the micropub.rocks permalink?
# 19:54 martymcguire[m] tantek: no real functionality changes, but the report format changed a bit and now it is verified with tests. :}
# 19:57 martymcguire[m] i've hit a point where, because i am working on two micropub clients with common server-side python functionality, i am trying to refactor into a library before adding more features. i should probably hold off on that and pass more client tests instead. :}
# 20:02 martymcguire[m] aaronpk: another question about working through the tests - there are tests for creating posts in both form-encoded and JSON format. is there a reason that a client would pass both of these tests?
KevinMarks joined the channel
# 20:06 bear IMO it should do both as being a client you don't know what the endpoint will accept
# 20:07 aaronpk You don't need to pass both if you have a reason to do one or the other only
# 20:07 aaronpk but keep in mind that some servers may only support form encoded "simple" posts
# 20:08 aaronpk ive been making Quill use form encoded for simple things and then upgrading to JSON requests when they use fancier features
[chrisaldrich] and gRegorLove joined the channel
# 20:14 martymcguire[m] is it goofy to ask the server what it supports? my reading of the spec suggests that a rule of thumb is "form encode until you want to do something that requires JSON"
# 20:17 bear IIRC that is why my client stuff does also
# 20:17 martymcguire[m] that brings me back to my confusion on the "200: Create an h-entry post (JSON)" vs "100: Create an h-entry post (form-encoded)"
# 20:18 martymcguire[m] following the "use form-encoding until you can't", my client would never do the 200 test without me telling it to, artificially
# 20:19 martymcguire[m] ok, cool. i will not let those unchecked boxes stress me out. :}
# 20:23 aaronpk I only implemented JSON responses for its token endpoint
KevinMarks and [dgold] joined the channel
KevinMarks joined the channel
# 21:55 ben_thatmustbeme hmm, interesting, aaronpk, so when i use json, i can upload media objects fine, but in form encoded mode, it get an error
# 21:55 ben_thatmustbeme photo%5B0%5D=https%3A%2F%2Fmicropub.rocks%2Fclient%2FOx5OgGCymkO6lkl5%2F700%2FRbYEGp08%2Ffile
# 21:57 aaronpk i can't remember the conversation that came up in, but we added this to the spec to clarify: "Syntax such as foo[0] and foo[bar] is not supported"
KevinMarks joined the channel
# 22:02 aaronpk hm yeah, but I should relax that on the server a bit
# 22:08 aaronpk oh wait, I think if you do upload something to the media endpoint it checks that off for you
KevinMarks joined the channel
# 22:10 aaronpk ah it's checking it off internally i just forgot to add the display
KartikPrabhu joined the channel
# 22:13 ben_thatmustbeme also, i take it back, i was able to pass some tests that i don't consider to be totally true
# 22:13 ben_thatmustbeme specifically, i can pass the tests that allow photo by url, just by using the media-endpoint
# 22:14 aaronpk the test doesn't require that the user enter an arbitrary URL
# 22:14 aaronpk it's about whether the server allows a URL for the photo property
# 22:15 aaronpk (ah I double checked and nevermind, I didn't actually finish implementing that checkbox, so it's still self-reported)
# 22:15 aaronpk yeah I don't want to dictate UI/UX stuff with the test suite, just checking protocol level stuff
KevinMarks and tantek joined the channel
# 23:05 Loqi Internationalization (AKA internationalisation, i18n; localization, l10n.) is the process of adapting software/content to various languages https://indieweb.org/i18n
# 23:05 Loqi Internationalization (AKA internationalisation, i18n; localization, l10n.) is the process of adapting software/content to various languages https://indieweb.org/l10n
KevinMarks and singpolyma joined the channel
KevinMarks_, gRegorLove and KevinMarks joined the channel