#dev 2017-10-25

2017-10-25 UTC
dougbeal|mb1 and leg joined the channel
#
aaronpk [eddie]: the job will run under whatever user you're logged in as when you're running the job
#
aaronpk so you'll probabyl have to use sudo to run it as the same user that fpm is running under
#
[eddie] ohhh okay. That makes sense
#
loqi.me edited /js;dr (+18) "tantek added "http://511.org/" to "See Also"" (view diff)
snarfed joined the channel
#
[eddie] !tell aaronpk: Victory! It works! My micropub server is upset about the json file, so I need to tweak that but besides that we've figured it out!
#
Loqi Ok, I'll tell them that when I see them next
#
aaronpk sweet!
[miklb] joined the channel
#
aaronpk oh yeah i think compass is still using a legacy combination of micropub+file upload that shouldn't actually be supported anymore
#
[miklb] spent the day at an intro to Google Cloud. Now I want to build *something*
#
aaronpk i shoulda thought of that before launching the GPS app
#
aaronpk https://github.com/aaronpk/Compass/issues/3
#
Loqi [aaronpk] #3 Update Micropub requests to JSON + Media Endpoint
#
loqi.me edited /js;dr (+41) "tantek added "http://511.org/accessible/tripplanner/" to "See Also"" (view diff)
#
[eddie] ohhhh lol
#
[eddie] That makes sense
snarfed joined the channel
#
bear raziellight - I use Kaku on my site and I run it using uwsgi and nginx, but it can run anywhere flask can. Please do ask me any questions you have
#
Loqi bear: tantek left you a message 23 hours, 34 minutes ago: for your review (off the top of my head because aaronpk asked) https://indieweb.org/email#phishing
snarfed joined the channel
#
bear the part that does the templating is distinct from the part that runs Flask. Flask is required only for the dynamic parts and uses all text files with a bit of redis for a simple event queue and state
EmreSokullu joined the channel
#
[eddie] aaronpk: Yeah, looking at my endpoint, I don't think I'd be able to add support for that in easily. So when I have the chance I might take stab at converting to proper media endpoint, etc instead of changing my endpoint to an unsupported feature. It looks like everything would be in the Jobs/TripComplete.php file that I would need to change. I can take a look at Quill for how you've done the media endpoints in the past
#
aaronpk yep that's the one
#
aaronpk basically it'd turn into a two-step process
snarfed1 joined the channel
#
aaronpk first uploading the JSON file to the media endpoint and storing the URL
#
aaronpk then changing the POST request in TripComplete.php to JSON format and using the URL of the file instead of adding the JSON data as a multipart chunk
snarfed, EmreSokullu and renem joined the channel
#
[eddie] That makes sense. I’ll take a swing at it when I get a chance.
tantek, EmreSokullu, [miklb], snarfed, cweiske and [eddie] joined the channel
#
[eddie] !tell aaronpk when trips are sent via Micropub do they cease to exist? (Wondering before I start making and testing Micropub changes with Compass and realize all my trips disappear ?) not the biggest deal because all of the data is still in compass but it’d be nice not to lose the trips.
#
Loqi Ok, I'll tell them that when I see them next
tantek, j12t, [chrisaldrich], EmreSokullu, [kevinmarks], calumryan, jeremycherfas and clintpatty joined the channel
#
loqi.me created /headless_CMS (+227) "prompted by sknebel and dfn added by sknebel" (view diff)
#
sknebel What is Netlify CMS?
#
Loqi Netlify CMS is a content management frontend for static sites stored in Github repos, made by Netlify https://indieweb.org/Netlify_CMS
#
sknebel Headless CMS << Example [[Netlify CMS]] - edits content stored on Github, actual website is rendered by a [[static site generator]]
#
Loqi ok, I added "Example [[Netlify CMS]] - edits content stored on Github, actual website is rendered by a [[static site generator]]" to the "See Also" section of /headless_CMS
#
loqi.me edited /headless_CMS (+135) "/* See Also */ new section" (view diff)
#
Zegnat cweiske, huh, apparently someone did bother to propose a way of linking to transcripts but not actually think of how to do transcripts themselves https://w3c.github.io/html-transcript/html-transcript-src.html
#
cweiske oh
#
cweiske thanks
#
jeremycherfas Impressive find.
#
Zegnat And, as linked from there, there are these things called text tracks: http://www.w3.org/TR/html5/embedded-content-0.html#text-track
#
Zegnat I am not familiar enough with them to comment on their use outside of media embeds though
#
Zegnat They can have cues for start and end times. But it is very much aimed at captioning and subtitling.
#
jeremycherfas Zegnat Clearly Tim Bunce has given a lot of thought to the mechanics of making transcripts, but I do wonder whether he has actually tried doing it as part of a podcast workflow. For some of mine, I have made the transcript, after the fact. And for some I have transcribed raw tape the better to edit it.
#
jeremycherfas I think I need to think again about how to do at least the timecoded transcript. There is a tool called F5 (made in Germany) that inserts timecodes automatically.
#
jeremycherfas Given a rough transcript produced by one of the online services, I think it would be reasonably quick to clean it up and insert timecodes with F5.
#
jeremycherfas Need to do the experiment.
#
Zegnat I have also seen people upload things to YouTube for their automatic captioning, and then download and work off of that.
#
Zegnat But I am looking from an HTML markup point of view only, and that seems very underdeveloped.
[kevinmarks] joined the channel
#
sknebel I wonder if "link to a subtitle file, or embed one in your audio file" doesn't achieve most goals
#
sknebel at least in regard to "provide metadata for external parties to handle"
#
cweiske yes, but not for web search
#
cweiske if you want to let people find the podcast by searching for keywords spoken in it
#
sknebel right. but does the text transcript the search engine sees have to be marked up specially?
#
cweiske no. but if you show the transcript on the page, it should be useful for the user - e.g. to let him jump to the specific position in the podcast by clicking on the line
#
sknebel I guess my question is, do the details/having a specific format matter that much if it is only used for page-internal functions like that. but I might have misinterpreted the initial question already
#
Zegnat We can't embed webvtt in html, can we?
#
cweiske it's plaintext like LRC
#
Zegnat Plaintext does not have to be a problem for search engines. Neither for the JavaScript you need to make the lines clickable.
#
Zegnat But for pointing the track element to it, I don't know.
j12t and jeremycherfas joined the channel
#
[kevinmarks] Well, that MP3 plus text track QuickTime movie is making my mac do rainbow spinnys
clintpatty joined the channel
#
jeremycherfas Zegnat, cweiske you know schmarty has been working on webvtt, I think.
clintpatty joined the channel
#
schmarty Thanks jeremycherfas, I should have been chatting here instead of the main channel
#
jeremycherfas Not at all.
clintpatty joined the channel
#
sknebel schmarty: btw, was it intentional that your last podcast tweet only had a picture and not a video? confused me quite a bit since I thought my browser was just to stupid to load the video in
#
schmarty checks
#
sknebel oh wait
#
sknebel it is a video
#
schmarty It plays for me
#
sknebel chrome took like 50s to load it to that point
#
schmarty Haha, it is a video with a still image and subtitles in
#
schmarty Oh wow
#
Loqi ahaha
#
schmarty It's only 65s long
#
sknebel yeah, I recognized the video template and was clicking it like mad. this time I left the tab open while chatting here and suddenly the video controls appeared
#
schmarty Hmm!
#
sknebel would be kind of nice if the first text line was visible in the poster image
#
sknebel then it's clear that it has subtitles
#
sknebel not sure how you can do that best in the video creation workflow
#
schmarty That's a good idea. I could probably get something in there.
#
jeremycherfas If I go back to my unmixed podcast audio, I can easily export separate files for each speaker, which could then be transcribed and merged back into a single transcript. No need to diarise the speakers.
clintpatty, raziellight, calumryan, EmreSokullu, eli_oat, dougbeal|imac, j12t, KartikPrabhu and [miklb] joined the channel
#
dgold how does one create an unauthorized access token for micropub.rocks 804?
#
aaronpk depends on how you issue access tokens
#
Loqi aaronpk: [eddie] left you a message 9 hours, 7 minutes ago: when trips are sent via Micropub do they cease to exist? (Wondering before I start making and testing Micropub changes with Compass and realize all my trips disappear ?) not the biggest deal because all of the data is still in compass but it’d be nice not to lose the trips.
KartikPrabhu joined the channel
#
aaronpk dgold: this tool might help https://gimme-a-token.5eb.nl/
#
dgold sudo gimme a token!
#
aaronpk get an access token without the "create" scope that way, and the test makes sure your endpoint rejects tokens without the "create" scope
#
aaronpk but if you store tokens in a database, or if you use JWT, you can create one manually as well
#
dgold I do neither! GDPR compliance is so much easier that way
#
aaronpk lol what
#
dgold thanks for that link, aaronpk
#
dgold means I can hit publish on this:- http://tinyurl.com/y86rp6cr [ https://micropub.rocks/implementation-reports/servers/132 ]
#
dgold oops - http://tinyurl.com/y8vpythx [ https://micropub.rocks/implementation-reports/servers/132/dohoQpnIdZYxrwcpMgzj ]
#
aaronpk nice!
clintpatty and EmreSokullu joined the channel
#
ben_thatmustbeme dgold++
#
Loqi dgold has 5 karma in this channel (23 overall)
#
dgold thank you ben!
EmreSokullu, John___, clintpatty, snarfed, tantek, j12t and Kaja_ joined the channel
#
loqi.me edited /headless_CMS (+72) "sknebel added "https://jamstack.org for the more generalized idea behind Netlify CMS" to "See Also"" (view diff)
#
sknebel What is headless CMS?
#
Loqi headless CMS is a Content Management System that only handles the backend parts of a traditional CMS (storage, editing interface, APIs, ...), with the reader-facing side handled by a different system https://indieweb.org/headless_CMS
#
sknebel KartikPrabhu: ^^^
#
aaronpk "reader-facing"?
#
aaronpk that's confusing
#
aaronpk not /reader I assume?
#
sknebel visitor-facing?
#
sknebel that better?
#
aaronpk yea
#
sknebel some people call it headend, which is where the headless comes from, but that was even more obscure
#
sknebel and frontend doesn't fit entirely, since the editing UI of the backend has a frontend too...
#
aaronpk i'm still kind of confused
EmreSokullu joined the channel
#
sknebel take a CMS, chop off everything a site visitor sees
[eddie] joined the channel
#
[eddie] In IndieWeb terminology a Headless CMS in the joining of a data storage engine and a micropub client
#
sknebel [eddie]: yes, pretty much
#
[eddie] They are mostly used by Javascript-powered sites and now, static sites are beginning to use them
#
aaronpk ah interesting
#
[eddie] They were initially created so that JS frameworks had an API they could plug into
#
sknebel and so you can have multiple frontends (e.g. website, mobile site, apps)
#
[eddie] That too
#
sknebel Netlify CMS is an example with a static gen. netlify hosts and builds static sites, so they made a editing UI that talks to the github API to generate the storage files
#
www.svenknebel.de edited /headless_CMS (+220) "better wording, example in indieweb tool terms" (view diff)
EmreSokullu joined the channel
#
GWG Howdy
#
sknebel hi GWG
#
GWG What goes on?
EmreSokullu and [manton] joined the channel
#
sknebel regarding the bandwidth discussion, interesting question/idea about forbidding non-gzip responses: https://news.ycombinator.com/item?id=15552973 (requires text-content to be the issue, and in the Indieweb world I wouldn't trust on every program fetching your site speaking gzip)
#
Loqi [jotto] Ask HN: Should we block non-gzip HTTP bots to reduce bandwidth costs?
#
aaronpk interesting
#
aaronpk that'd certainly help with tantek's atom feed
#
aaronpk in my case the bandwidth usage is all images
#
sknebel it would probably "help" by breaking a bunch of feed readers
#
sknebel I would assume things that understand gzip responses also announce that, but I guess one or two popular libraries could have forgotten that
EmreSokullu joined the channel
#
www.boffosocko.com edited /Mastodon (+1481) "POSSE; IndieWeb federation, Fed.Brid.gy" (view diff)
#
raziellight can anyone point me in the right direction for token and auth end point implementations. i'm trying to understand how it all works and trying to connect with a client
#
raziellight i'm using python btw
#
aaronpk https://indieweb.org/authorization-endpoint and https://indieweb.org/token-endpoint
#
aaronpk and this is a good walkthrough https://indieweb.org/obtaining-an-access-token
#
raziellight ya read those
#
sknebel aaronpk: unless I'm completely missing something, tantek's site doesn't even deliver his atom feed gzipped at all
#
raziellight ya i saw that
#
aaronpk sknebel: that seems quite possible. that should normally be handled at the web server level so i wonder why
#
sknebel could have been missed in the config, since generally the file types for which it is done have to be set
#
raziellight i'm going to read this again.
#
sknebel I feel like those pages miss some implementations listed, but I can't remember any specific one
[chrisaldrich] joined the channel
#
sknebel !tell tantek: one thing I noticed regarding your traffic issues is that apparently only static files get gzip-compressed by your webserver. maybe they let you configure that?
#
Loqi Ok, I'll tell them that when I see them next
[cleverdevil] and EmreSokullu joined the channel
#
dgold aaronpk++ got switchboard integration working on my site - thanks for the clarification
#
Loqi aaronpk has 80 karma in this channel (1453 overall)
#
aaronpk oh fun!
EmreSokullu joined the channel
#
dgold was just glad it takes a json-payload for hub.mode & hub.url
#
aaronpk it does?
#
dgold well, I got a 200
#
aaronpk interesting. i bet the framework i'm using converts that automatically
#
dgold time to write some itches on my wikipage
#
ascraeus.org edited /User:Ascraeus.org (+567) "/* My Indieweb Setup */" (view diff)
#
ascraeus.org edited /User:Ascraeus.org (+58) "/* nanopub */" (view diff)
gRegorLove, tantek, EmreSokullu, [miklb] and [dgold] joined the channel