#dev 2021-05-12

2021-05-12 UTC
[tw2113_Slack_] joined the channel
#
@MMontevil
I've redone my #academic website using @eleven_ty It is fun to play with various data sources like @CrossrefOrg api as a complement to #webmention. I also took some inspiration and code from @nhoizey https://montevil.org
(twitter.com/_/status/1392291491448492041)
sparseMatrix joined the channel
#
sparseMatrix
whew I finally got my webserver up and running again, this time /with/ ssl ala let's encrypt, and on nginx....should be a lot less trouble to support than lighttpd.
#
sparseMatrix
tommorow I'll get my h-card+blog back up and running
#
sparseMatrix
...and, get back to development on my microformat/micropub project
ShadowKyogre, Seirdy, gRegorLove, [tantek], jacky and [chee] joined the channel; ShadowKyogre left the channel
#
[chee]
what do people do for pictures that aren't "photo", like illustrations and other visual artworks? kind of post, microformats etc. lmk if this is the wrong channel for this question
gRegorLove joined the channel
#
[chee]
it seems strange to me that "photo" is the kind of image post. there is "audio" and "video", which are like... abstract, non-specific. and then "photo", which excludes illustrations and other designs
#
[chee]
i guess i should just post them as notes?
#
sknebel
[chee]: its afaik more that photo was the primary use case people thought about, but any post that's specifically the image can be a "photo" post, even though it is a bit misnamed for that
#
[chee]
i'm going to use notes with tags instead. the naming is unfortunate, and i want to distinguish between photos and other visual art. it's not quite satisfying to have a `photo` category with a sub-type of `photo` for actual photos, so i will use tagged notes for non-photo visual art
[dianoetic_net] and [KevinMarks] joined the channel
#
sknebel
I mean how you call the thing in your UI is unrelated to the microformats
#
sknebel
wait, "kind of post", are you talking wordpress post kinds?
#
sknebel
ah yeah
#
sknebel
yeah, thats just how it evolved, not naming you need to use if you dont like it
#
[chee]
it does say, "photo, or other image" on that page but the words "or other image" are vastly outnumbered by the words that assume "photo" means "photo" lol
#
sknebel
yes because thats the most common example. but again, just ignore that when thinking about non-photo images, not like you need to use photo-specific details for that
#
sknebel
(and e.g. my site doesn't expose what "post type" something is (doesnt even have a concept for it), so for visitors it really doesnt matter)
[dianoetic_net] joined the channel
#
[dianoetic_net]
Is there a conventional name for Atom feeds?
#
doosboox
[dianoetic_net]: atom.xml?
#
doosboox
or feed.xml
Ethan[m], ShadowKyogre and gxt joined the channel; ShadowKyogre left the channel
#
[dianoetic_net]
Thanks! Just wondered what was typical
[KevinMarks] joined the channel
#
@indigitalcolor
↩️ If the error may be related to third-party look at its source code. I’ve looked at React, misc. packages with types, Webmention, etc on GitHub to better understand their functionality and find relevant open issues. Often times others may have encountered a similar issue.
(twitter.com/_/status/1392464061099565058)
[Murray] joined the channel
#
[Murray]
[dianoetic_net] if you're thinking terms of feed discovery, the /feeds idea may be useful: https://marcus.io/blog/making-rss-more-visible-again-with-slash-feeds
#
Loqi
Making RSS more visible again with a /feeds page
#
[Murray]
(I'm sure this is documented on the Wiki somewhere but I cannot find it 😅 )
alex11 and sparseMatrix joined the channel
#
sparseMatrix
wow, nginx reverse proxies are a blast (if you're into that kinda thing)
#
@lcptuk
Webmention W3C protocol for notifying a URL when a website links to it https://en.wikipedia.org/wiki/Webmention
(twitter.com/_/status/1392480028747776009)
[dianoetic_net] joined the channel
#
[dianoetic_net]
Thanks [Murray] ! Also Loqi 😄
#
Loqi
you're welcome, [dianoetic_net]
#
[Murray]
you're welcome too 😄
ShadowKyogre and [KevinMarks] joined the channel; ShadowKyogre left the channel
ShadowKyogre, [KevinMarks]1 and [tantek] joined the channel; ShadowKyogre left the channel
#
[tantek]
[dianoetic_net] I use "updates.atom" named after the user-centric meaning (It's a bunch of recent "updates") and the extension .atom because Atom does have a specific MIME type to serve it rather than generic text/xml or application/xml
#
[tantek]
I'm not a fan of meaningless or redundant names like "feed.xml"
#
[tantek]
Hmm, I read [Jeremy_Keith]'s post about /feeds and while I respect that it's "easy" to setup and may help with /discovery for some folks, I see no reason to promote or otherwise try to (re)make RSS "a thing".
#
[tantek]
it's got a plumbing-centric smell IMO
#
[tantek]
also propagates the whole notion of feeds being a separate thing with its own meaning rather than "just" another format of the thing you're viewing in HTML
#
[tantek]
AKA the thing you're viewing in the browser
[Jeremy_Keith] joined the channel
#
[Jeremy_Keith]
RSS *is* a thing. People are reading RSS feeds in RSS readers. Yes, it’s plumbing. So is every format. As plumbing goes it’s not the worst.
#
aaronpk
i would argue as plumbing goes, it's one of the worst
#
aaronpk
JSONFeed is better RSS plumbing than RSS
#
[tantek]
There's a trade-off between resurrecting an approach that only ever appealed to a very small minority, and doing the harder work of pushing for a consistent user-centric approach that appeals to (and is this approachable / accessible to) a much broader set of folks.
#
[tantek]
Or to put it another way, sure, we can fetishize the RSS ecosystem of the early/mid-2000s, or we can acknowledge the UX improvements & innovations from social media that have made it easier for more people to write and read each others writing
#
[Jeremy_Keith]
And note that the URL is /feeds, not /rss …that would be plumbling-centric.
#
aaronpk
where does the name "feed" come from anyway
[KevinMarks] joined the channel
#
aaronpk
is it from printers?
#
[Jeremy_Keith]
And I think RSS has a bit of UX advantage over JSON in that it can be styled (admittedly with quite a bit of hoop-jumping on the publisher’s end …but it’s a benefit for the user). e.g. https://adactio.com/rss
#
[Jeremy_Keith]
I think the feed terminology is more from news rooms. Publications would subscribe to, for example, the feed from Reuters. Ticker tape back in the day, I think.
#
[tantek]
anything where you make clickable links that return unstyled windowful of XML soup is a bad UX
[jacky] joined the channel
#
[tantek]
CDF mentioned "updated channel" but not feed
#
[jacky]
Wow, that's something I learned today (the history of feeds)
#
[Jeremy_Keith]
“anything where you make clickable links that return unstyled windowful of XML soup is a bad UX” That’s my point: because the RSS is styleable, the user doesn’t see the XML.
#
aaronpk
i used to love the idea of XSLT back in the day
#
aaronpk
until i had to go build anything with it 😂
#
[jacky]
I was going to ask if someone had a guide for XSLT
#
[Jeremy_Keith]
Granted, most sites don’t style their RSS. I wish they would.
#
aaronpk
wonders if someone has made JSLT yet
#
[jacky]
takes away [aaronpk]'s keyboard for 5 minutes
#
[jacky]
ah okay thanks [Jeremy_Keith]
#
[tantek]
[Jeremy_Keith] when I clicked on https://adactio.com/journal/tags/sci-fi/rss from your /feeds page I got the unstyled XML soup
#
[Jeremy_Keith]
Really? That’s odd.
#
aaronpk
i see a nice page there in chrome
#
[Jeremy_Keith]
I see a nice page there in Firefox
#
[jacky]
hm, it's styled for me (that link [tantek] shared)
#
[Jeremy_Keith]
I think the XSLT transformations depend on JavaScript.
#
aaronpk
in some ways, XSLT is the ultimate separation between data and presentation layers
#
[Jeremy_Keith]
It’s like JSSS all over again (who remembers that?)
#
[tantek]
ah looks like NOSCRIPT was blocking XSLT 🤦‍♂️
#
[jacky]
tbh this is something I'd like to have in my feeds - I could even put a little blurb
#
[jacky]
wants to add this to the Wiki
#
[jacky]
found another post actually _from_ the Wiki
#
Loqi
[Hsiaoming Yang] How to style RSS feed
[Ana_Rodrigues] joined the channel
#
[Ana_Rodrigues]
I didn’t even know that was possible (styling RSS feeds) 😮
#
[jacky]
Heh, hidden nuggets
#
[tantek]
right, it's a convoluted way to reinvent HTML+CSS, when there was no need to.
#
[tantek]
or to be less generous, it's echoes of an attempt to reinvent the web in XML which died in the early 2000s
#
[tantek]
like a historical curiosity where the reward is discovering the dead-end of a long abandoned effort
#
aaronpk
didn't firefox have some default styling for RSS feeds a while ago?
#
[tantek]
got removed at some point
#
[KevinMarks]
have you considered h-feed -radical idea, but it's a feed in HTML so it can be styled naturally
[tw2113_Slack_] joined the channel
#
[tw2113_Slack_]
i see https://cloudup.com/cZZ_PPDKqZj when viewing directly
#
[KevinMarks]
also firefox styles json now
#
[Jeremy_Keith]
“have you considered h-feed -radical idea, but it’s a feed in HTML so it can be styled naturally” Why not both?
#
doosboox
I’ve considered making my post archive page an h-feed
#
[Jeremy_Keith]
It’s not HTML vs. RSS. It’s HTML + RSS (+ JSON …+a Twitter feed …+ however else someone wants to get at my content)
#
[jacky]
granted, I think my personal choice to go Atom whenever I can (mainly from https://en.wikipedia.org/wiki/RSS#RSS_compared_with_Atom) might help with a bit of future-proofing for me
#
[tantek]
Atom has a well defined spec with processes for updating it (IETF). RSS is a dead-end with ambiguities that will never be resolved (except by whatever dominant implementations choose that year).
#
[Jeremy_Keith]
I tend to use the label RSS to mean “either RSS or Atom”
#
[tantek]
here I thought that's what you used the "feed" label to mean
#
[tantek]
hence /feeds instead of /rss
#
[Jeremy_Keith]
“feed” covers even more I think: there’s also JSON feeds on offer (and Twitter feeds).
#
[jacky]
yeahs tbh
#
[tantek]
feels like unnecessary hierarchy
#
[Jeremy_Keith]
But I might tweak the copy on my /feeds URL to remove most of the instances of the word “RSS” at all. Even if it is RSS under the hood, that doesn’t matter to the subscriber.
#
[tantek]
it shouldn't matter yeah
#
[Jeremy_Keith]
No hierarchy: RSS, JSON, and Twitter are all equal—different formats but all of them are feeds that can be subscribed to.
#
[tantek]
presumably Atom would be in that list then
#
[tantek]
rather than being a nested child of "RSS"
#
aaronpk
XML :)
#
aaronpk
call it what it is :D
#
[KevinMarks]
I could add links via granary to create the other kinds I suppose
#
[tantek]
right, JSON is not a feed. JSONFeed is
#
[KevinMarks]
or write some code instead of my manual until it hurts homepage
#
[jacky]
*yeah tbh I think I could see me doing that for my /follow page re: naming it
#
[Jeremy_Keith]
Right, I’ve updated the copy here: https://adactio.com/feeds/
#
[Jeremy_Keith]
I’ve also aliased any urls that end with /feed to /rss
#
[KevinMarks]
typo insterested
#
[tantek]
jacky, a /follow page makes more sense to me for several reasons
#
[tantek]
focused on user intention (action) rather than jargon
#
[tantek]
using language more people already understand (have been trained by social media), rather than essentially a throwback
#
[KevinMarks]
so is https://podcastsconnect.apple.com the worst feed reader ever?
#
[KevinMarks]
I've been trying to fix an archived podcast feed for it, and each update takes a day to return
[manton] joined the channel
#
[manton]
[KevinMarks] The rollout of the new Podcasts Connect has been pretty rocky. I’ve hit a couple bugs and also heard from other folks about feeds not updating right away. I get the impression they had to rush out the new stuff to support paid audio subscriptions.
#
[KevinMarks]
it's been very flaky for me
[aciccarello] joined the channel
#
[aciccarello]
I've thought about making a `/subscribe` page that would link to my atom feed and the eventual email newsletter format I'm trying to setup from my feed.
[dianoetic_net] and shoesNsocks joined the channel
#
[tantek]
worth experimenting with that [aciccarello] to see how it feels once built, and to see what kind of folks end up using it (hopefully successfully)
#
[tantek]
people also "subscribe" to podcasts right? rather than "follow" a podcast?
#
aaronpk
subscribe to podcasts, subscribe on youtube
#
[aciccarello]
Yeah. "Subscribe" does have more of a channel connotation.
#
[dianoetic_net]
I know how to create a feed with a URL that ends in `atom.xml` or something. How does it work if I want to make `dianoetic.net/subscribe`? Does the page at that address just need to contain links to the `atom.xml` one?
#
[aciccarello]
[dianoetic_net] I'm thinking something similar to https://adactio.com/feeds/
#
[aciccarello]
Just different path idea + email newsletter form
#
Loqi
Jeremy Keith
#
[manton]
Apple switched to “Follow” instead of “Subscribe” in the latest version of their Podcasts app. I think the idea is that “Subscribe” will be used for paid shows. Hard to tell yet if that trend will be adopted for podcasts elsewhere, though.
#
[dianoetic_net]
That's a fair point. "Subscribe" could imply payment
#
petermolnar
that is not a nice direction for wording. If I'm not paying, I'm a "follower" (I still dislike the religious/cult ringing to it), if I do, I'm a "subscriber" - to me, that sounds quite bad, like the non-payers are mere sheep, following someone.
#
[KevinMarks]
lol, it wants exactly 3000 by 3000 pixel jpeg for logo art
#
petermolnar
sounds like someone is about to do 4K favicons
#
[KevinMarks]
well, it gave me a url for the podcast in their index, but it's not there https://podcasts.apple.com/us/podcast/tummelvision/id1566968860
#
[KevinMarks]
is this backend still running on the old infrastructure? How can it take days to update?
lermis joined the channel
#
aaronpk
i see the podcast but no episodes
#
aaronpk
the podcasts app on my mac just says it can't connect
barnabywalters joined the channel
#
[tantek]
maybe that's for AppleTV? (the large logo art)
#
[KevinMarks]
spiderpig made the original feed into an index.html because it don't have an extension, so I've been hand editing it into compliance
#
[KevinMarks]
anyway, I'll check tomorrow as their update cycle is slow AF
#
[Jeremy_Keith]
Right, I’ve updated https://adactio.com/feeds/ so there is no only one instance of the phrase “RSS” (even where RSS is indeed the plumbling under the hood)
#
Loqi
Jeremy Keith
#
[tantek]
[Jeremy_Keith]++ that makes sense. Presumably you're looking at https://adactio.com/feeds/subscriptions/ next?
#
Loqi
[Jeremy_Keith] has 1 karma in this channel over the last year (5 in all channels)
#
Loqi
Jeremy Keith
#
[tantek]
also I checked an RSS0.9 also doesn't mention "feed". I got those links from /timeline
#
[tantek]
which begs the question, what was the first "feed" spec that actually mentioned "feed" explicitly?
#
petermolnar
the wordpress feed is under example.com/feed/
#
petermolnar
maybe we should simply follow that
#
petermolnar
(the main feed, that is)
#
petermolnar
[jacky]: re XSLT before you start going down the same rabbithole as I did: there is NO way to turn CDATA into "real" HTML, because the function is not implemented in every browser - for example in Firefox. So the only thing one can show is a thin feed (link, title, pubdate), but not a full-fledged page, which includes the description.
#
petermolnar
so yeah, following WordPress' lead, my feed is /feed/, more specifically /feed/index.xml
#
[dianoetic_net]
Works for me ¯\_(ツ)_/¯
#
[tantek]
I just didn't think a feed/feeds merited a whole folder/directory but that's me
#
[tantek]
that is /feed/index.xml feels even more characters for no reason than /feed.xml
#
superkuh
I thought wordpress had it at /feed not feed/
#
superkuh
When manually checking I normally go through, /feed , /index.xml, /rss.xml
#
[tantek]
do we need manual checking? feed discovery via rel=alternate is fairly well supported
#
superkuh
Mozilla removed support for showing feeds in their browser. People have much less incentive to do rel=alternate these days with browsers ignoring feeds.
#
[tantek]
the two are unrelated
#
[tantek]
rel=alternate to feeds was almost always on invisible <link> tags
#
[KevinMarks]
The advantage of feed.xml is for static sites to serve the right mine type
#
[tantek]
nothing to do with what's shown in the browser or not
#
[tantek]
KevinMarks, that's wrong too, as ".xml" won't give you the right MIME type
#
[tantek]
this is precisely why I use ".atom" not ".xml"
#
doosboox
[tantek]: I thought an atom feed was xml?
#
[dianoetic_net]
Good to know!
#
doosboox
Oh! It's something like text/xml+atom isn't it?
#
[tantek]
doosboox, Atom has its own MIME type
#
[tantek]
application/ but yeah
#
doosboox
ah! ty
#
doosboox
yeah, my apache gives application/xml for my feed
#
[dianoetic_net]
Oh good I have that
#
[tantek]
"application/atom+xml"
#
doosboox
I'd have to add a new mime type to apache for that
#
[tantek]
put "AddType 'application/atom+xml' .atom" as a line in your .htaccess
#
[tantek]
then the ".atom" file naming/handling will take care of it
#
doosboox
[tantek]: yeah... But I'd also have to rename my atom.xml file :D That'd ruin it for anyone who's already following it
#
aaronpk
redirect?
#
doosboox
or symlink, I guess
#
doosboox
eh. I'll put it on my todo list and see if I have an opinion when I get to it
#
doosboox
:D
[jacky] joined the channel
#
[tantek]
is now deep diving on his htacsess which he clearly hasn't touched in >2 years.
#
barnabywalters
last time I checked, HTML5 Boilerplate was a really good source of htaccess tips and optimisations
#
barnabywalters
I think mine is based on some version of that from many years ago
#
petermolnar
heh. nginx doesn't have htaccess, so I have an absolute abomination php that handles all 404-s :D
#
[KevinMarks]
I made a whole ass appengine app to redirect one domain
[jeremycherfas] joined the channel
#
petermolnar
that's ... decication
#
petermolnar
*dedication
#
[tantek]
desication, as in drying out a dead thing in order to preserve it, kinda makes sense there too
jacky joined the channel
#
[tantek]
[KevinMarks] have you seen this before? I just discovered it myself: https://wiki.mozilla.org/Firefox/Feature_Brainstorming:Notes_and_Annotations
#
[KevinMarks]
That comes around every so often as an app idea. Remember Google sidewiki?
#
[tantek]
I mean it's a page from 2007 about some of the concepts we've discussed here and I'd never heard of it before
#
[tantek]
it goes beyond any one app idea
[Murray] joined the channel
#
[Murray]
Woops, didn't mean to kick off a whole conversation about /feeds 😄
#
[Murray]
FWIW though, RE: RSS being "bad", I see RSS as much more useful than stuff like JSON feed (and even h-feed) simply because an ecosystem exists to consume/support it (Atom lumped into RSS here, though I honestly have no idea if I've ever come across Atom in the wild, wouldn't know the difference). I know social readers exist, but I've never understood the appeal (for my personal use, I get why others would use them and think they're generally a
#
[Murray]
cool idea), and there doesn't seem to be very many. I like TheOldReader, it works well, its easy to setup and manage, and any of my non-techie friends that became interested have liked it (or similar sites, like Feedly). RSS is just easy, and most sites I've looked at support it; heck, I've even contacted sites that didn't and they've got it running extremely quickly.
[fluffy] joined the channel
#
[fluffy]
yeah like I agree with the notion that h-feed makes for a nicer mode of presenting things (especially if <link rel=“feed”> is supported) but all the tooling right now is around rss/atom. Also rss/atom are a lot easier to parse without relying on an existing library, although really you should be parsing with a library.
#
[fluffy]
For the foreseeable future it seems like everything (both readers and publishers) should support rss/atom.
#
[fluffy]
Just like how websites still need to support favicon.ico in addition to <link rel=“shortcut icon”> or whatever.
#
aaronpk
good luck parsing rss/atom without a library heh
#
aaronpk
i've tried a number of times to use just an XML parser and it's not fun at all
#
aaronpk
h-feed doesn't really have a better story there though. even if you make the analogy of XML Parser :: Microformats Parser actually consuming the parsed result of both is still a lot of work
#
[KevinMarks]
you really need a library for rss/atom. if you don't believe that try the 3500 test universal feedparser has
#
[jacky]
parsing is hard
#
[tantek]
I'd say that understates the true difficulty of "parsing" RSS. Like you can't "just" treat it as XML and have it work with the existing "ecosystem" because there are so many examples of weirdly broken.
#
[tantek]
oh whoops didn't see aaronpk, [KevinMarks] responses. yeah what they said
#
[tantek]
business model << Some documentation & criticism of content creation / subscriptions / patronage as business models: 2021-05-08 The Economist: [https://www.economist.com/briefing/2021/05/08/the-new-rules-of-the-creator-economy The new rules of the “creator economy”] / Social-media platforms used to get most of their content for free. That dynamic is changing
#
Loqi
ok, I added "Some documentation & criticism of content creation / subscriptions / patronage as business models: 2021-05-08 The Economist: [https://www.economist.com/briefing/2021/05/08/the-new-rules-of-the-creator-economy The new rules of the “creator economy”] / Social-media platforms used to get most of their content for free. That dynamic is changing" to the "See Also" section of /business-models https://indieweb.org/wiki/index.php?diff=75664&oldid=73496
#
@ChrisAldrich
↩️ I post on my site first (frequently using Micropub clients) & syndicate to Twitter (known as POSSE), usually using http://Brid.gy Publish. I also use http://Brid.gy for backfeed to get the responses back via Webmention. https://boffosocko.com/2021/05/12/55791046/?replytocom=321610#respond
(twitter.com/_/status/1392627537805643778)
#
@ChrisAldrich
↩️ I post on my site first (frequently using Micropub clients) & syndicate to Twitter (known as POSSE), usually using http://Brid.gy Publish. I also use http://Brid.gy for backfeed to get the responses back via Webmention. https://boffosocko.com/2021/05/12/55791046/?replytocom=321610#respond
(twitter.com/_/status/1392627537805643778)