#dev 2017-07-05

2017-07-05 UTC
[miklb], zoglesby, AngeloGladding, snarfed and [acegiak_net] joined the channel
#
@WendyandCharles ReadersGazette: BLOG Indie Author Answers by Jim Heskett http://www.thejugglingauthor.com/indieauth/ Get help writing your book #bookbloggers 8 (twitter.com/_/status/882447250844262400)
myfreeweb, schmarty, snarfed, tantek, [miklb], barpthewire, cweiske, [kevinmarks] and j12t joined the channel
#
sknebel hm, I can't get backtitude to work :/
#
sknebel it thinks I have no location providers
barpthewire and j12t joined the channel
#
Zegnat I believe it is working for me. I just haven't written the custom endpoint for it to sync to yet. So I can't be sure what data it captured.
#
petermolnar sknebel in case of android 6+ go into settings/apps/backtitude and check the permissions
#
petermolnar Zegnat you need to configure it to save locally into kml and csv files
#
petermolnar if you don't do that, it doesn't collect data
#
petermolnar Backtitude -> Settings -> Internal Memory Storage Options -> select KML and CSV
#
sebsel To speak with GWG loosely: what is it about location that keeps people interested? :)
#
petermolnar I'm tracking that to geotag my photos
#
sebsel I feel like cheating because of the location chat of you guys I now use aaronpk's app and (Apple fanboy alert) it just works :')
#
petermolnar backtitude works perfectly for me
#
sebsel yeah, that's where I want to go to, geotag posts (for myself)
#
petermolnar I even wrote an endpoint for it which also worked, but realized the KML files make more sense to me
#
petermolnar I have one more reason though: the first time we were in China I wasn't tracking my location well enough, and the next time we went there we were unable to find a store we really wanted to, which we visited the first time
petermolnar joined the channel
#
sebsel yeah! I get that. Since IWC Nürnberg I have (OwnYour)Swarm, and I noticed I was checking in to all kinds of places because I liked the location archive.
#
sebsel So now I have a non-public, non-silo, non-web way of doing that.
#
petermolnar I really don't like the idea of publicly sharing I'm not at home
#
petermolnar if someone pokes me, sure
#
Zegnat petermolnar, i
#
Zegnat I am storing both kml and csv
#
Zegnat Stupid phone stupid enter
#
petermolnar in that case: /sdcard0/Backtitude/[files] - open them with some text reader, eg. get Ghost Commander
#
sknebel petermolnar: android 4.1 :/
#
sknebel sebsel: travelling right now, and want afterwards a list of places I was and geotagged photos
#
sebsel I'm not doubting use cases, it's just interesting to see that it keeps interesting people :)
#
sknebel sebsel: didn't take it as doubt, wanted to provide info :)
j12t and davidmead joined the channel
#
Zegnat petermolnar, I feel very little for emailing myself those files every now and then ;) Which is why I am looking into how Backtitude is able to POST the information to an endpoint
#
petermolnar I can help with that later
#
petermolnar I've done it once already
#
petermolnar fairly straightforward
#
thejeshgn.com about created /User:Thejeshgn.com_about (+227) "Created page with " [Thejesh GN (ತೇಜೇಶ್ ಜಿ.ಎನ್)](https://thejeshgn.com) “Thej” is an Independent Technologist, Hacker, Maker, Traveler, Blogger, InfoActivist, Open data an..."" (view diff)
#
@thej IndieAuth https://indieauth.com/ is very interesting. (twitter.com/_/status/882570544666759169)
tantek, barpthewire, j12t, eli_oat, eli_oat1, [jeremycherfas], singpolyma, dougbeal|mb1 and snarfed joined the channel
#
barryfrost.com edited /User:Barryfrost.com (+511) "/* IndieWeb projects */ Adding my recent projects" (view diff)
drewinthehead joined the channel
#
drewinthehead webmentions implementation: is there any reason not to have a unique endpoint URL per content item? Spec seems to allow for query params, so I think that should be ok. Anything I'm overlooking?
#
aaronpk a few of us were doing that for a while
#
drewinthehead I sense a but...
#
ben_thatmustbeme whats the reasoning for it though?
#
aaronpk well specifically some of us were doing expiring endpoints
#
ben_thatmustbeme its definitely allowed, and been used
#
aaronpk i ended up dropping that idea after running it for 6 months and looking at the results
#
aaronpk also some notes here: https://indieweb.org/Webmention-brainstorming#checking_endpoint_once_per_domain.2Fcaching_the_endpoint
#
drewinthehead the reasoning is that the implementation is sufficiently generic that it's hard to determine the content item ('post') from the URL alone. I deliver the content, but don't necessarily control the URL handling.
#
aaronpk not sure i follow why it's hard to determine the post from the URL
#
aaronpk you get a target parameter which is the URL of the thing the person links to. you're saying it's hard to look up what that URL is in your backend?
#
drewinthehead I'm saying I don't know anything about the URL and how the user is working with it to determine what content might be on the page.
#
aaronpk oh are you making a plugin or some sort of middleware thing?
#
drewinthehead Just a regular CMS
#
drewinthehead User builds a site. We provide page functions they can use in a toolkit style to pull out content where they want it.
#
drewinthehead Best I could do is request the local page myself and look for clues. (Or plant clues)
#
aaronpk interesting, i still don't quite understand why, but i at least see the problem
#
drewinthehead so I'd like to use a unique endpoint that has the post ID as a param
#
aaronpk you can definitely do that. the only caveat is that some senders (i think only bridgy right now) will cache the webmention endpoint they find for a site. i'm not totally sure on exactly what bridgy caches tho, e.g. if it's per URL or per domain
#
drewinthehead I guess that's what I'm not clear on - what's the scope of an endpoint? to me the spec currently reads like the endpoint has page scope
#
drewinthehead "endpoint for a site" seems like an implementation assumption
#
drewinthehead I think I'll give this a go. If I need to rethink further down the line, I can discard the extra detail without problems.
#
aaronpk yeah the intent of the spec is that the endpoint is per URL
#
aaronpk however in practice bridgy had to do some caching, so that's an interesting data point
tantek joined the channel
#
drewinthehead cool, that's what I thought. thanks aaronpk
tantek and snarfed joined the channel
#
@rubygems webmention (0.1.6): A gem for sending webmention (and pingback) notifications https://rubygems.org/gems/webmention (twitter.com/_/status/882634295214575617)
#
thejeshgn.com created /User:Thejeshgn.com (+339) "Created page with "[Thejesh GN (ತೇಜೇಶ್ ಜಿ.ಎನ್)](https://thejeshgn.com) “Thej” is an Independent Technologist, Hacker, Maker, Traveler, Blogger, InfoActivist, Open data and..."" (view diff)
#
thejeshgn.com edited /User:Thejeshgn.com (-2) (view diff)
#
thejeshgn.com edited /User:Thejeshgn.com (+3) (view diff)
#
tantek wonders how bad it would be if he synchronously serverside called (curled) stream thatmustbe us and webmention io for every event permalink rendering
#
tantek is tempted to try it until something breaks or someone complains
#
tantek alternatively, I wonder if a "stylesheet=" param on stream thatmustbe us could work well enough to provide an inline stylesheet that it would return in a <style> element, to embed the result in an iframe
#
tantek decides to file the feature request to find out
#
tantek continues to go down the display-rsvps-as-a-service rabbithole
#
Zegnat Now that scoped style got scrapped, I am not sure I would be comfortable pulling in external CSS, tantek
eli_oat joined the channel
#
tantek Zegnat, not a problem for iframe
#
tantek if I was pulling in external HTML, I would do the CSS myself
#
Zegnat Ah, yes, within iframes it should be safe enough
#
tantek my first "how bad it would be" question is about pulling in external HTML serverside, which is likely slower and more fragile, so I'm leaning against trying it, to avoid setting a bad example
snarfed joined the channel
#
tantek filed: https://github.com/dissolve/socialstream/issues/10
#
Loqi [tantek] #10 add optional style= param, use to output <link>/<style> stylesheet when ashtml=1
#
tantek is really enjoying the small pieces loosely joined of tantek.com -> webmention discovery -> webmention.io -> query jf2 result -> stream.thatmustbe.us -> HTML+mf2 and soon: + style -> iframe embed RSVPs on event permalink page
#
aaronpk i'm seriously considering adding an embed option to webmention.io to make it easier than using the JSON API
#
aaronpk but i'm not sure which approach is better... providing an iframe option, or providing an option where you can fetch the HTML from JS and embed it into the pagee yourself, which would let you control the styles more
#
tantek aaronpk - you see where I'm going right?
#
tantek iframe is better. works in noJS :D
#
tantek how is that even a question
#
tantek :)
#
tantek you can control the styles just fine with an iframe option, see https://github.com/dissolve/socialstream/issues/10 for how
#
Loqi [tantek] #10 add optional style= param, use to output <link>/<style> stylesheet when ashtml=1
#
aaronpk cause not everone cares about that :P
#
aaronpk they're both js;dr for scrapers
#
tantek except archive org will index an iframe
#
aaronpk wait iframe has a style option?
#
tantek what's the scraper use case for responses
#
tantek aaronpk no just read https://github.com/dissolve/socialstream/issues/10
#
Loqi [tantek] #10 add optional style= param, use to output <link>/<style> stylesheet when ashtml=1
#
aaronpk oh i see
#
aaronpk there's not much of a scraper use case for that, which is my point, that using client-side JS isn't a huge drawback
#
tantek everytime you can avoid using JS, you make an experience more secure, more robust
#
tantek also, better automatic caching behavior by the browser
#
tantek seriously, technically this should not ever be a question
#
tantek "not everyone cares about that" is subject to cargo-cult idiocy
#
aaronpk has the iframe height problem been solved?
#
tantek I'll find out when there's something to prototype with with actual content that has an actual height that I care about
#
tantek (as opposed to all theoretically possible heights)
#
aaronpk this is not a theoretical problem
#
tantek note that the RSVPs section on FB is of fixed height and that's not a problem.
#
tantek on FB events
#
tantek so yes, I'm claiming this is theoretical for this use-case until we have prototypes to try out
#
aaronpk that's a clever way to avoid the problem
#
tantek which is *why* we prototype first, worry about theoretical problems later
#
tantek it may (often) turn(s) out that problem space you care about is disjoint with the theoretical problems that people raise who have yet to implement the feature on their own site
Jeena joined the channel
#
tantek that reminds me ...
#
tantek ... in #indie-chat
#
tantek er #indieweb-chat
#
snarfed intresting, new jetpack release adds support for liking comments in wordpress. https://jetpack.com/2017/07/05/jetpack-5-1-comment-likes/
#
Loqi [Richard] Jetpack 5.1: Comment Likes and Enhancements
#
tantek what is a reply chain
#
Loqi A reply chain (AKA reply thread) is a list of replies, and replies to those replies, displayed under the original post https://indieweb.org/reply-chain
#
loqi.me edited /tweetstorm (+152) "tantek added "2017-07-04 [https://twitter.com/NPR/status/882313133863436288 @NPR tweetstorms the US Declaration of Independence] (100+ tweets long [[reply-chain]])" to "See Also"" (view diff)
snarfed and snarfed1 joined the channel
#
ben_thatmustbeme heh, saw your issue on github tantek, i knew right away that was where you were going with it
#
tantek ben_thatmustbeme: I realized that iframes have much better caching / network efficiency behavior
#
tantek :D
#
ben_thatmustbeme http://stream.thatmustbe.us/?url=https%3A%2F%2Fwebmention.io%2Fapi%2Flinks.jf2%3Ftarget%3Dhttp%3A%2F%2Ftantek.com%2F2017%2F183%2Fe1%2Fsunday-yoga-presidio+&op=jf2-mf2&ashtml=1&style=body%7Bcolor%3Agreen%7D
#
ben_thatmustbeme it works
#
tantek this is too cool
#
tantek proceds to code CSS in the address bar
#
ben_thatmustbeme lol
#
ben_thatmustbeme it takes URLs as well
#
ben_thatmustbeme probably a better option
#
ben_thatmustbeme no promises about security with that code though
#
ben_thatmustbeme well, rather i'm quite sure its insecure as you could to style=</style><badcode<style>
#
tantek ben_thatmustbeme: alright that's fixable, one sec
#
tantek digs through his XHTML archives
#
tantek s/<style>/<style>/*<![CDATA[*/
#
tantek s/</style>//*]]>*/</style>
#
tantek (that's for your own generation code)
#
tantek and then just one escape regex to apply to the style param value
#
tantek s/</&lt;
#
tantek may cause invalid stylesheets but that's not a problem for CSS parsers which are good at handling invalid input due to the test suite
#
tantek ok I need the p-rsvp in a data element, and the value to be an actual p-rsvp value instead of a URL lol
#
tantek but that's a webmention issue
#
tantek er, webmention.io issue
snarfed joined the channel
#
ben_thatmustbeme that doesn't seem to fix anything tantek
#
@rubygems jekyll-webmention_io (2.3.0): This Gem includes a suite of tools for managing webmentions in Jekyll: * Tags -… https://rubygems.org/gems/jekyll-webmention_io (twitter.com/_/status/882668099211862016)
#
tantek ben_thatmustbeme: huh? by escaping any "<" on input, you avoid any tag embedding via the style param
#
ben_thatmustbeme http://stream.thatmustbe.us/?url=https%3A%2F%2Fwebmention.io%2Fapi%2Flinks.jf2%3Ftarget%3Dhttp%3A%2F%2Ftantek.com%2F2017%2F183%2Fe1%2Fsunday-yoga-presidio+&op=jf2-mf2&ashtml=1&style=body{color%3Agreen}%3C/style%3E%3Cscript%3Ealert(%27hello%27)%3C/script%3E%3Cstyle%3E
#
ben_thatmustbeme oh, i missed that
#
ben_thatmustbeme huh
#
tantek yes that's kinda key
#
tantek the CDATA helps with some degree of tag defensiveness (from XHTML/XML days)
#
ben_thatmustbeme okay, thats all fixed
#
tantek next issue filed
#
tantek https://github.com/dissolve/socialstream/issues/11
#
Loqi [tantek] #11 better markup for jf2>mf2 ashtml p-rsvp values though data elements
#
tantek and then I'm going to quickly run into webmention.io bugs
#
tantek aaronpk, re: "seriously considering ... webmention.io to make it easier" - these smaller fixes would make it easier: https://github.com/aaronpk/webmention.io/issues/81 and especially https://github.com/aaronpk/webmention.io/issues/80 !
#
Loqi [sknebel] #81 inconsistent handling of upper/lowercase in RSVP processing
#
tantek Loqi can you quote the other one for me too? https://github.com/aaronpk/webmention.io/issues/80
#
Loqi [sknebel] #80 JF2 API: rsvp property overwritten by URL
#
tantek continues to hack address bar CSS as much as he can to get this closer
#
aaronpk k!
eli_oat, [miklb] and [pfefferle] joined the channel
#
tantek ben_thatmustbeme: slight improvement tweak on socialstream #9 https://github.com/dissolve/socialstream/issues/9#issuecomment-313194057
#
Loqi [tantek] Per this example: http://stream.thatmustbe.us/?url=https%3A%2F%2Fwebmention.io%2Fapi%2Flinks.jf2%3Ftarget%3Dhttp%3A%2F%2Ftantek.com%2F2017%2F183%2Fe1%2Fsunday-yoga-presidio+&op=jf2-mf2&ashtml=1&style=img{height%3A44px}.h-feed%3E.p-name,.h-entry%3E.u-...
#
tantek also heads-up aaronpk I'm going to be feature requesting aggregate results for webmention links
#
aaronpk aggregate results?
#
tantek you know, for things like "7 Went • 1 Interested"
#
tantek purely (not) hypothetically speaking ;)
#
ben_thatmustbeme tantek, just pushed some additional updates to socialstream
arush1 joined the channel
#
tantek aggregate results: https://github.com/aaronpk/webmention.io/issues/82
#
Loqi [tantek] #82 Feature request: aggregate webmentions for a link (# of rsvps of each type, # comments / reposts / likes)
[kevinmarks] joined the channel
#
tantek ben_thatmustbeme: got it. that helps!
#
tantek is getting really close to the limitations of this approach without further webmention.io fixes :)
#
[miklb] fixes or features?
#
tantek [miklb]: fixes specifically issues 80 and 81
#
[miklb] just curious.
j12t joined the channel
#
tantek and limitation hit, though I'm pretty pleased with the progress so far towards showing an RSVP facepile. example at the top of https://github.com/dissolve/socialstream/issues/12 if you're curious (let's see how Loqi does in summarizing that ;) )
#
Loqi [tantek] #12 use data element for p-wm-* properties and put them first in h-entry
#
tantek oh right, Loqi doesn't dig into the contents there. phew :)
#
tantek wonders if sorting is better done before jf2 output or in jf2->HTML+mf2 conversion
gRegorLove and snarfed joined the channel
#
tantek decides JSON because JSON arrays are ordered by dfn
#
tantek decides jf2 that is
#
tantek ok I *think* I'm done filing webmention.io feature requests (hopefully as small enough bitesize incremental improvement chunks)
#
tantek (at least for this particular display RSVPs use-case)
snarfed, j12t and gRegorLove joined the channel
#
gregorlove.com edited /Homebrew_Website_Club (+548) "/* Regular Meetings */ venue consistency notes" (view diff)
#
petermolnar is anyone aware of CMS-es which, by default, include PESOS of anything?
#
snarfed if RSS counts, then lots
#
tantek snarfed, I would not count RSS by itself as that is just plumbing
#
tantek if a CMS had a feature like, "auto-post your Flickr favorites as likes on your own site", THAT would be PESOS by default, regardless of what mechanism it used under the covers, whether Flickr favorites RSS feed, Atom feed, proprietary API etc.
#
snarfed sure, RSS is mostly plumbing, but it leaks into UX pretty heavily. iirc micro.blug, tumblr, wordpress can all automatically pull another site in and PESOS its posts, ongoing
#
tantek I know it's a fine distinction to draw, but I think the UX/user-centric distinction is quite important
#
snarfed and i think they all say, "this works via RSS" in the UI
#
tantek certainly for anything considered "by default"
#
snarfed ie the feature isn't tied to a specific silo via API
#
tantek IMO "by default" for a user means 1. you don't have to know anything about any plumbing, and 2. you certainly do not have to go dig for an RSS feed somewhere and then figure out a form field UI to paste various things into and enable
#
tantek snarfed, it may be worth documenting screenshots of where such plumbing leaks into UX, as ideally it should not. Exposing users to technical jargon typically makes sites and features less accessible, not more.
#
tantek "this works via open standards" is good. "this works via RSS" is not.
#
tantek there is a big difference there
#
snarfed sure. you're assuming more than i said. i think those examples are all the former. they're built in, and i believe generally do feed discovery. they probably do mention RSS, but i expect it's downplayed
#
snarfed agreed, documenting would be great!
#
tantek I'm probably filtering out such jargon in the UX, which is why I suggested screenshots if you're seeing "this works via RSS" in the UI
#
petermolnar RSS is not PESOS
#
petermolnar it only goes back X posts
#
petermolnar not all of them
#
tantek petermolnar: PESOS I think implies future, not past
#
tantek you're talking about backfilling
#
tantek which is something completely different and can be done with an import / export etc.
#
petermolnar yes and no
#
petermolnar backfilling, yes; import, in a way
[miklb] joined the channel
#
tantek though there are exceptions like Swarm/Foursquare .ics feed has *all* of your "posts" (checkins)
#
petermolnar but if you think about it there's nothing stopping PESOS from "importing" old entries
#
tantek sorry, but that's literally not what PESOS stands for
#
tantek it's a process, and starts with *post elsewhere* as a present verb, not past
#
petermolnar that's a nice touch from 4⧠
#
tantek that's a very deliberate definition to clarify the "live" usage
#
petermolnar it'll always be a backfilling
#
petermolnar post elsehwere, pull to your site
#
petermolnar the only difference is a timestamp
#
tantek while technically correct in terms of non-zero time, that's certainly not the user-perceived meaning
#
tantek no, that is not the only difference
#
tantek from a user action perspective, nothing happens *until* they post
#
petermolnar I disagree here, to be honest
#
tantek *elsewhere*, and then in a good system, nearly instantly they see it on their site
#
petermolnar I'd expect my content to appear
#
petermolnar even the old ones
#
tantek PESOS is literally about user-behavior, not systems behavior
#
tantek I think that may be the source of confusion
#
tantek you can *manually* PESOS for example
#
petermolnar tantek there's nothing in /PESOS that make a distint difference between these behaviours
#
petermolnar and it's not a confusion
#
tantek as above, it's the literal definition using a present tense verb
#
petermolnar I'd expect a pesos plugin to import the old things
#
tantek instead of past tense or entirety
#
tantek so yes the definition makes this clear
#
petermolnar maybe not by default, but it should be able to do it
#
tantek you just want more than PESOS
j12t joined the channel
#
tantek you want PESOS + backfill
#
petermolnar which is pesos...
#
tantek no because backfilling is not associated with current posting action
#
petermolnar "publishing flow starts with posting to 3rd party services, then using some infrastructure (e.g. feeds, pingbacks, webhooks) to create an archive copy under your domain. "
#
tantek and it can be done as a separate batch operation etc.
#
petermolnar posted in 2014
#
petermolnar now I want it on my site
#
petermolnar that definition does not contradict it
#
tantek that's stretching the use of a present tense verb
#
tantek by a few years :P
#
tantek your conflation of a 2014 action with the present is as silly as saying "I am attending the 2014 IndieWebCamp in New York City"
#
petermolnar that's a time limited action
#
tantek not a useful use of a present tense verb
#
petermolnar whereas posting an artwork, for example, is not
#
tantek in both cases you finished the action, it is in the past
#
tantek unless you are taking years to paint a painting
snarfed joined the channel
#
tantek This distinction is useful because PESOS has some realtime expectations, whereas backfilling does not
#
petermolnar I'm not going to be convinced on this: yes, it may be stretching the language, but putting a time limit on PESOS makes no sense; the archival copy on you site will always be after the canonical in the silos, which makes PESOS a method, of, by definition, capable of backfilling.
#
tantek right, I'm saying you're missing the point of user expectations
#
tantek posting means it shows up, not it shows up years later
#
petermolnar did you ask users about these expectations,
#
petermolnar ?
#
tantek and yeah if it takes more than a few seconds, it's annoying, if it takes more than a few minutes, it seems broken
#
tantek indeed, these came out actual selfdogfood discussions regarding POSSE
#
petermolnar and if my pesos is running from a 10m cronjob?
#
aaronpk OwnYourSwarm has exponential backoff of polling times
#
aaronpk at best it will be a few seconds, worst will be a day
#
tantek if you built it then that is the expectation, and you won't be surprised
#
aaronpk same with OwnYourGram
#
tantek aaronpk: do they document the latencies anywhere?
#
aaronpk http://ownyourgram.com/docs
#
petermolnar "that is the expectation" -> this is what I'm talking about, that my expectation is that a PESOS plugin would also backfill on first run
#
aaronpk importing past content is a whole different story. often times peoples' sites aren't even ready to handle that, but they still want to support PESOS from the point of signup on ownyourgram
#
tantek right
#
aaronpk the best thing for me personally is realizing that importing past content doesn't have to be tied to importing future conrtent
#
aaronpk all of a sudden I started getting my instagram photos and swarm checkins on my site. i'll deal with importing the old stuff later
#
aaronpk having to wait to be able to support importing the past content is a huge barrier and would have taken me a lot longer before i started owning that data
#
petermolnar but what if you have a wordpress, which can easily backfill anything; a pesos plugin, which can access all ancient and future data - would you not expect a question asked "shall I import old stuff as well?"
#
tantek petermolnar, yes, numerous users agree on this use of backfill, see: https://www.google.com/search?q=site%3Aindieweb.org+backfill and in general https://www.google.com/search?q=site:indieweb.org+PESOS as well
#
tantek sounds like a plugin that has both PESOS and backfill features
#
tantek and sure, you can design a plugin like that. there's just no need to conflate the two features, which is only confusing.
#
tantek backfill itself means importing past posts
#
aaronpk my point is there's value in separating the ideas
#
tantek typically since the beginning
#
tantek aaronpk: indeed, my point as well. diferences in latency expectations, realtime vs batch etc.
#
petermolnar I get the naming. I get that there's value in separating. But there is also value in merging the true, especially for newcomers, who have a lot of data in silos and they want it out
#
petermolnar nobody wants to deal with parsing a fb export
#
gRegorLove What is backfill?
#
Loqi It looks like we don't have a page for "backfill" yet. Would you like to create it?
#
gRegorLove What is import?
#
Loqi It looks like we don't have a page for "import" yet. Would you like to create it?
#
tantek what is export
#
Loqi export (specifically, How to export your data) in the context of the indieweb refers to the ability to download some or all of your data from a site, typically from a silo, though also for CMS or site migration https://indieweb.org/export
#
tantek import is export
#
Loqi ok
#
loqi.me created /import (+44) "prompted by gRegorLove and dfn added by tantek" (view diff)
#
tantek :P
#
petermolnar Loqi tantek is bending the truth
#
gRegorLove wat
#
petermolnar import != export
#
aaronpk lol
#
tantek backfill is the action of importing all your past posts, typically from a social media silo, into your own site.
#
loqi.me created /backfill (+139) "prompted by gRegorLove and dfn added by tantek" (view diff)
#
Loqi ok
#
petermolnar anyway, in case of CMS plugins I think backfill should be a question asked when someone is setting a PESOS plugin up, since the code is most probably there to pull the data in
#
tantek I would not make any such assumptions
#
tantek and in fact would assert the opposite
#
tantek based the evidence we have from existing PESOS services
#
tantek which do *future* posts (and maybe posts from today or your most recent post) only, and would take A LOT more work to crawl through all your past posts
#
petermolnar and you'd add +1 plugin to gen4 users?
#
aaronpk i haven't dug into it, but i'm pretty sure i'm gonna have to go through hoops to find past instagram photos. right now I use an endpoint that returns the last N photos from the user. there is no clear paging mechanism though
#
aaronpk not to mention if this is going to be self-hosted somewhere, setting up a mechanism to crawl a ton of old posts has requirements that fetching future posts does not
#
tantek no I'd rather ship a good plugin sooner for gen1-2 users rather than never ship a perfect plugin to gen1-4 users
#
aaronpk (background jobs, queue management, showing import progress)
#
petermolnar I was able to fetch every favs/likes from DA, Tumblr, Flickr and 500px, all paginated at some point, but no problems
#
petermolnar I'm still not using instagram, so that I don't know
#
petermolnar I'm guessing pulling all the things from facebook is near impossible
#
petermolnar tantek you have a point there
#
tantek petermolnar++ for being able to fetch all of anything from any silo - that is very hard work IMO
#
Loqi petermolnar has 4 karma in this channel (38 overall)
#
petermolnar only favs/likes
#
petermolnar and it wasn't that hard to be honest
#
petermolnar they all have their own little quirks, which would be lovely to have standardized ('paged', 'has_more', 'offset'... )
#
petermolnar and no, not everything, I'm wrong there
#
petermolnar 500px changed it's system a while ago and completely removed any way of listing what you liked
#
petermolnar they used to have separate like and fav
#
petermolnar now it's like or add to gallery
#
petermolnar see https://support.500px.com/hc/en-us/articles/236067827-Where-can-I-view-my-Likes-
#
petermolnar anyway, this was a good discussion
#
petermolnar thanks tantek and aaronpk
#
tantek even just favs/likes backfill would be an amazing service!
#
tantek what is 500px
#
Loqi 500px is a photo-sharing silo https://indieweb.org/500px
#
tantek what is DeviantArt
#
Loqi deviantART is an art hosting silo that retains the rights to resell copyright of anything you post to it, without informing you, and has done so in the past https://indieweb.org/deviantART
#
tantek And I know /Tumblr /Flickr exist already
#
petermolnar how would you do a service like that?
#
tantek petermolnar: can you document what you got working in terms of "fetch every favs/likes" from each of those in a "Backfill" section?
#
tantek on each of those service's pages?
#
petermolnar ugh
#
petermolnar I'll try to
#
tantek would be great to capture that, even if it's just open source that you ran yourself on your own server
#
petermolnar that'll take a while
#
tantek pretty sure there are many others that would benefit and want to do so too
#
tantek even just "{{petermolnar}} used project XYZ to backfill all his likes/faves from his silo account to his own site"
#
tantek whatever the XYZ project is, even if it is not open source, at least document that you were able to do it
#
tantek hopefully just a two line wiki edit for each of those pages
#
petermolnar I have the code on github, opened, though calling it open source is a stretch, and I need to merge a few things which is lacking at the moment
#
tantek feel free to caveat that too! that's ok to totally be transparent about current status of the code
#
tantek sometimes just seeing it documented that it is possible provides hope, and that alone is useful
[manton] joined the channel
#
[manton] aaronpk I updated Micro.blog's Micropub support today so that OwnYourGram works for hosted sites. Pretty cool to see it work! (Previously I required clients to use the media endpoint, which doesn't look like OwnYourGram checks for. Also added a rel="me" for Instagram.)
#
Loqi [manton]: [eddie] left you a message 3 weeks, 4 days ago: you note permalinks have the date/time link within the h-entry content. I'm on mobile right now so I can't look at your source code for suggestions but the mf2 if definitely a little off
#
aaronpk ooh nice
#
aaronpk OwnYourGram doesn't upload to the user's media endpoint, but if it detects one when you sign in it will switch to sending JSON posts with the Instagram URL as the photo parameter, instead of uploading the file
#
aaronpk (you can change that in the "settings" section on the dashbaord after you log in)
#
[manton] Cool, I see that option. Multipart seems to be working fine now, though.
#
[miklb] manton while you’re around, can you point me to best place to ask about using micropub with a self hosted WordPress.
#
[miklb] on micro.blog iOS app I mean
#
petermolnar tantek https://github.com/petermolnar/nasg/blob/master/pesos.py - I do need to document it, but just in case, that's the main worker at the moment
#
petermolnar plus that thin oauth.py next to it
#
[manton] miklb Right now the iOS app always defaults to XML-RPC for WordPress, actually, and only switches to Micropub if it can't discover the endpoint. I've been thinking about splitting that out into a setting so that you can force it to always use Micropub.
#
petermolnar if I want to make my code useful I need to remove a lot of hardcoded things from the config
#
petermolnar including getting userids where needed
[cleverdevil] joined the channel
#
[manton] miklb You might be able to force it to use Micropub if there's no "EditURI" link tag in your web site's response. Also, FYI, I just fixed a Micropub client bug that was effecting some servers and it's waiting for Apple to approve it.
KartikPrabhu and j12t joined the channel
#
GWG I am still hoping for a native Micropub client for Android
#
[miklb] manton thanks! I thought for some reason WP turned off xml-rpc by default, but now thinking about it, was only for a short time
#
[cleverdevil] Hmm, I need to connect my Grill's microblog to Micro.blog.
#
[manton] miklb Yeah, might've been disabled a while back, but now it's on by default. Actually I don't think there's a setting to toggle it off, so you'd have to block xmlrpc.php or maybe find a plugin to disable it?
#
[cleverdevil] ?
#
[manton] cleverdevil Yes! ?
#
[cleverdevil] Can haz another invite?
#
[manton] Sure, I'll send you one.
#
[cleverdevil] Thanks ?
#
[cleverdevil] I will have the world's first device connected to Micro.blog!
#
[cleverdevil] History in the making.
#
[miklb] looks like there’s a filter to return false to turn off xml-rpc
#
[miklb] I will try it. Thanks again!
#
[manton] miklb Cool, let me know if it works! (Also, very likely you will need to wait until version 1.0.1 is approved by Apple. There was an encoding bug with some of the parameters.)
#
[miklb] gotcha.
gRegorLove and snarfed joined the channel
#
tantek !tell aaronpk,ben_thatmustbeme with the fixing/implementing for socialstream issues 11 & 12, and webmention.io issues 80 (and preferably 84) I'm pretty sure I'll have what I need create an RSVPs facepile embed very similar to a popular silo's RSVPs facepile. Going to have time to code on this in the next few days.
#
Loqi Ok, I'll tell them that when I see them next
j12t, billbennettnz and snarfed joined the channel