ZegnatIt is interesting how basically all search results I get for “piwigo” and “export” are about tools that help you take your photos from other software *into* piwigo
sknebel[snarfed]: but the two webmentions kaushalmodi got only differ in the twitter-URL, the bridgy markup doesn't show anything about it being a retweet
kaushalmodiI am confused.. It's bridgy that sets post vs repost in the URL, right.. I am just receiving WM's from webmention.io which sets the activity type as "link" for both: original and retweet (in this particular case)
ZegnatInteresting, wagle. I’d say a lot of people here are fans of pulling things out of databases and putting it in plain text files for long term storage of data.
ZegnatBut yeah, I am not seeing a lot of export options for Piwigo. Maybe if you look for other open-source photo tools that they will have made “import from Piwigo” tools. Otherwise it looks pretty locked up
ZegnatNote that I am in the annoying position of not being able to pronounce my own name very well, else I would have recorded a clip of my entire name.
schmartybecause in the sense that tons of human effort to get a universal textual representation of mouth sounds, it feels like the "right way to do it"
aaronpkI think an audio file is the best option, since it removes the need to find common written symbols to represent different sounds from different languages
aaronpkCross language "baby talk" spelling doesn't work very well either cause there are many sounds in other languages that don't exist in English and can't be spelled with English words
schmartywikipedia/wiktionary are good places to check. my previous strategies have involved trying to find videos where a person is mentioned or says their name, or appearances on podcasts.
aaronpkZegnat: in that case I would say the more important thing is that people are pronouncing your name the way you want them to whether or not it's "technically correct"
schmartymy current strategy is usually to check their site's about info to figure out where they live (or if they say where they are "from"), or the primary non-english language that they post in on their site
Zegnataaronpk, problem comes from not being able to say that. I wonder for instance how Jonathan Ross pronounces his own surname. And if that becomes “Woss”, does he actually want people to call him that?
ZegnatI too sometimes struggle with Rs, and hearing people try to parrot the way I pronounce my own given name, with them trying to mimic my R, just sounds wrong.
[jgmac1106]This will happen often with different languages, especially if moving from a phonetic based reading language to one that isn't but even then there are simply some phonemes (a human invention) that are used in Hindi and English vice versa that our mouths are not trained to make
jmacAnd working all these together as a little IPA teaching-nudge, hmm? e.g. <p class="p-phonetic-name"><a href="http://example.com/me-pronouncing-my-name.mp3" class="p-ipa u-sound">[My name in IPA characters]</a></p> (Just spitballing here)
Zegnatp-ipa might be a good idea. Would set it aside from other pronunciation guides, and as a mf2 consumer you know you only have to look in there if you are able to use IPA in the first place
schmartyjmac: i think u-sound, being derived from hCard's "sound" and vCard's "sound" before that, would be understood to be a spoken sound clip of the person or organization or place's name
ZegnatYou could document that. But the raw microformats aren’t going to show a relationship between properties. All properties are descriptors of the upper object. So all mf2 shows is that `sound` and `ipa` are descriptors of `h-card`.
schmartyjmac: oh, good question. i don't believe there is, but i'd be interested in using it to link podcast episodes to their transcriptions, or link the this week in the indieweb newsletters with their audio editions.
jmacZegnat: But if a u-sound in an h-card is already understood to mean "This is the 'name' of this h-card read out loud", then that's all we need; the fact that you can represent that as a hyperlink from IPA text is a happy accident.
schmartyzegnat: did you mean to write "not (yet) part of the official h-review spec" on the mf2 wiki regarding u-sound? seems like that should be h-card, but i get easily confused about the mf2 wiki content. ;}
[kevinmarks]I'd say avoid nesting properties - we flattened h-card to make it easier. If you want pronunciations of other things (organisatiosn etc) then embed an h-card for them wiht it's own u-sound
ZegnatLooks like rel="alternate" is page scoped. So can only use that for the audio version links on permalink pages and not within feeds. Am I right in that reading of the HTML spec, [kevinmarks]?
ZegnatThe question is what people will understand the sound property to be ... if they expect to find the full post in audio format, 10 seconds may be a let down. More a [.u-summary .u-sound] then, maybe?
Zegnatwagle, probably because you are free to take all your data. As all the data is indeed stored by you on hosting you pay for. They just aren’t telling you exactly you can access that data ;)
jmacYour IPA looks like your name, mine looks like horrible shoggoth-speak. I'm trusting that my language-nerd friends in another chatroom aren't playing a joke on me.
Loqi[Kaushal] Finding some really interesting static-site pipelines. My favorite so far is @kaushalmodi's: gitlab.com/kaushalmodi/ka…
I wish there was a mor
jmacZegnat: A valid question, and for the time being I would rather stay respectfully mum on this topic! (Which is an entirely separate crate of potatoes I intend to write up later.)
Loqi[John Feminella ⌬] Finding some really interesting static-site pipelines. My favorite so far is @kaushalmodi's: gitlab.com/kaushalmodi/ka…
I wish there was a more general effort to making pipelines agnostic of what they were running against.
snarfedkaushalmodi: no, i haven't changed any code. bridgy just renders a retweet to html a bit differently depending on whether you explicitly tell it it's a retweet or not
[kevinmarks]that thing where none of your folders have spaces in because python install tools fail unpredictably when they do, and every python programmer has already done this so it never gets fixed…
bdeshamOr could I return something like urn:uuid:f81d4fae-7dec-11d0-a765-00a0c91e6bf6 and only "promote" the file to a "real" URL once a post was created that actually referenced the file?
ZegnatGood question. From spec as written, any valid URL pointing “to the file” should be a valid response. Though I think the intent was for it to be resolvable.
bdeshamYeah, I figured that was the case. It seems preferable to me, though, to make the uploaded file publicly-accessible only when the relevant post is published, instead of making it accessible immediately and then expiring it at some point in the future
aaronpkadvantage of it being accessible immediately: see the quill UI for adding a photo to a post. it shows the post as a preview before the note is created
ZegnatI guess some editors may have issue with it if they try to display an uploaded image inline? Does the Quill article editor not do something like that?
bdeshamaaronpk: I haven't uploaded a photo with Quill yet (although I have used it for several text posts--thank you very much for making it!). Does Quill upload the file and then download it again to create the preview?
bdeshamI was thinking that since the URL returned from the media endpoint is (potentially) temporary anyway, returning a non-resolvable URI wouldn't be too bad. But it sounds like that would break some clients
Loqiurn is a Uniform Resource Name that comes from the original design of Internet resources but is often now a source of dead links https://indieweb.org/urn
LoqiAn urn is used to store the ashes of cremated people, however URN is an abbreviation for Uniform Resource Name, specified as a kind of URI, yet later abandoned by the URL Standard, and historically often a source of dead links, ironically aligned with the original common noun.</span> https://indieweb.org/urn
singpolymaHow can a urn be a "dead link"? Most of them are stuff like urn:isbn:... ISBNs don't expire or anything, and in terms of resolution are never "live links" to begin with
singpolymaI mean, we could deband what "standards body" mean all day I suppose :P I'm probably just too used to open stardards, but arguably ISO and everyone are just as closed as whatwg so maybe none of it means anything :P
singpolymaThis list of 10 things is good, but except for indieweb/microformats does anyone operate like this? IETF is close, except for 8,9,10. Other groups tend to fail much earlier (due to not being open in at lest of of develop/iterate/discuss)
bearmy original point, before the rabbit holes and such, was that defining URN purely from a web standard view isn't going to give the proper depth of a definition - but i'm going to stop because i'm tired of being told to not worry about anything that isn't a pure web experience
tanteksingpolyma: if you want to drop "It's 2018" type statements, then yes, it's 2018 and no standards group/body that calls itself that has any excuse for not doing open (*free*) test suite and open implementation reports - it was a common best practice 7 years ago as I blogged it
singpolymaI'm in the awkward situation where I need to go for unrelated reasons, but if I disconnect it'll look like I'm leaving a semi-heated discussion in a huff. So this is just to say I have no ill feelings and am leaving for unrelated reasons :) Big respect for you tantek and aaronpk as always
tanteksknebel: that daniel haxx se link also notes: "The only somewhat modern "spec" for URLs is the WHATWG URL specification. The other major, but now somewhat aged, URL spec is RFC 3986, made by the IETF and published in 2005." (with links)
@rombulowYou know how HTTP GET requests are meant to be idempotent? Well, do I have the story for you ... a while back I added WiFi control to our garage doors with little Wemos D1s. (twitter.com/_/status/990684453734203392)
LoqiAn RSVP is a reply to an event post that says whether the sender is or is not attending, might attend, or is just interested in the event https://indieweb.org/multirsvp
snarfed, jgmac1106, [kevinmarks], tantek, KartikPrabhu and leg joined the channel; bdesham_afk left the channel