#dev 2022-04-06

2022-04-06 UTC
darkkirb, cygnoir[d], Darius_Dunlap[d], Natlogica[d], nertzy, sarahd[d], jacky and angelo joined the channel
#
chewing_bever
so i've just noticed that the web server that i use has support for serving the compressed version of a file if it's present and the client sends the correct Accept-Encoding header
#
chewing_bever
would it be worth the trouble to implement this?
jamietanna1 and chewing_bever joined the channel
#
[KevinMarks]
You can measure it with lighthouse. It can help for low bandwidth connections
[James_Van_Dyne], [KevinMarks], nsh- and jan6 joined the channel
#
capjamesg[d]
search << There are many people and businesses working on new takes on search that contain new ideas: https://dkb.io/post/the-next-google
#
Loqi
ok, I added "There are many people and businesses working on new takes on search that contain new ideas: https://dkb.io/post/the-next-google" to the "See Also" section of /search https://indieweb.org/wiki/index.php?diff=80641&oldid=80637
tetov-irc, johnnrs[d], nertzy, jacky, Silicon[d], aspenmayer[d], Darius_Dunlap[d], MarkJR84[d], petermolnar, grantcodes[d], keenan[d], capjamesg[d], Christian_Olivie, sayanarijit[d], rattroupe[d], cygnoir[d], dovedozen[d], Jeremiah[d], sarahd[d], mossymaker[d], Eddy04[d], samhenrigold[d], YimingWu[d], hoenir, marksuth[d], Natlogica[d] and toyinariyo joined the channel
jacky joined the channel
#
capjamesg[d]
Is that layout confusing?
#
jacky
a little bit
#
jacky
is the part with "Cinnamon" bigger than the body text on purpose? if so, what does it represent?
hans63us[d] joined the channel
#
jamietanna
I'd probably expect the reader view (left) to be bigger than the response view (?, right)
#
jacky
^ yeah
#
[manton]
Good point yesterday [tantek] about the safety / owning your content appeal of static files. That is kind of what I was curious about… How can we help people feel they have control of their files without the technical hurdles? Something I think about with Micro.blog export and syncing options.
#
jacky
syncing feels like the most common option
#
jacky
a personal hope is to get some sort of peering of information happening so it's like a personal 'torrent' of your site's data
#
jacky
but perhaps the other side of this is needed - a way to import and restore from said backups (even if it's like demostrative)
#
jacky
I know [manton] you proposed a format for these backups
#
[manton]
I bet this is also why some people like Blot, which just uses Dropbox files.
#
jacky
what is blot.im
#
Loqi
Blot is an open source flat file blogging platform built by David Merfield that uses Dropbox, Google Drive, or Git to store your content https://indieweb.org/blot.im
#
[manton]
[jacky] Yes! I’ve done more recently with the format… I’d love for more people to support it. (It’s basically a ZIP file with Microformats and JSON Feed and images.)
#
[manton]
What is blog archive format?
#
Loqi
blog archive format is a data format proposed by Manton Reece for the export of a blog, based on a zip file and top level HTML h-feed inside, that is supported by micro.blog https://indieweb.org/blog_archive_format
#
jacky
yeah, I think I'm going to look into supporting this format
#
[manton]
Cool, let me know if you have any questions.
#
[manton]
I think there’s a place for both “just a folder of static files” and “easy ZIP file that has everything in one place”. The latter is convenient for backups and importing.
#
jacky
that's something I'm slowly working on with my site
#
[manton]
Here’s a longer overview with context that I wrote for my book: https://book.micro.blog/blog-archive-format/
#
jacky
it has an abstraction on most of its file operations (trying to move theming and SQLite to be over this adapter as well) but it could point to something living in S3 or any particular location
#
jacky
oh nice
niklasfyi[d], jacky and edburns[d] joined the channel
#
[tantek]1
[manton] exactly. there's nothing like being able to browse a folder of static plain-ish text-ish (actual plain text, "markdown" pretending to be plain text, or even HTML)
#
[tantek]1
that behavior illustrates that people do care about /longevity even if they don't explicitly identify it as such
#
[tantek]1
if you can't open a file in a (plain) text editor and read at least most of the content, then the file format sucks (for text). images, video, audio don't have an equivalent
[fluffy] joined the channel
#
[fluffy]
Yeah this is why Publ uses a folder of Markdown files as its ground-truth storage and only uses the database to index that folder for content retrieval. All rendered text comes from the ground-truth file.
#
[fluffy]
all the configuration is stored in the Publ site’s folder too, like it’s in the form of a Python virtual environment that’s ready to be reconstituted and deployed
#
[fluffy]
so like if you keep it in a git repo you can also (relatively) easily back out config changes too
#
jacky
I think that's what I'm going to sync my setup over
#
jacky
over Git
#
[tantek]1
taking the Twitter editing your website issue here for more dev-specific discussions (JS) etc. and strategies
#
[tantek]1
what is embed
#
Loqi
embed is code (usually just HTML like an iframe, sometimes with JS) for showing content from another website on your website or inside a post that is included by viewer’s browser and typically has some interactive aspect (more than just a static image / audio / video file) https://indieweb.org/embed
#
[tantek]1
^ [KevinMarks] I feel that page could be improved a bunch to document the Twitter example / misbehavior, and strategies for how to do better, both as a publisher embedding stuff, and as a service providing stuff to be embedded
#
[KevinMarks]
I'll have a look. I think Ben's solution was pretty good at the time (when twitter was 140 characters) but lost fidelity as they added more.
#
[tantek]1
I did a bit of minor clean-up to make some space for you 🙂
#
[aciccarello]
[jacky] What are you thinking for post type linting? I've thought about having some unit tests that parse's my site's rendered HTML to make sure the microformats don't change.
#
jamietanna
I used to do it with the HTML-Proofer Ruby gem that ran against the generated HTML - definitely was useful to catch things when doing bit theme changes
#
Loqi
[edent] SuperTinyIcons: Under 1KB each! Super Tiny Icons are miniscule SVG versions of your favourite website and app logos
[denschub] joined the channel
#
[denschub]
po
#
[denschub]
↩️ ehm - sorry, that was a copy-paste error that somehow ended up wrong. I also can’t delete this message, so… have fun with it!
#
jacky
so I'd like it to be something I can run against a Web page
#
jacky
but also in my unit tests
#
[aciccarello]
Cool, FYI the link to what you wrote on your site isn't working for me
#
jacky
yeah that's in the ether now (lol)
#
jacky
I still have the data but I have yet to restore into my new site
#
[aciccarello]
:headstone:
#
[aciccarello]
For your unit tests would you just check the parse microformats for an expected structure or should there be some shared library or something that is more of a semantic representation.
#
jacky
heh yeah
#
jacky
I'd be parsing for an expected structure (does it show tags, syndication info, etc)
#
jacky
(I have a map of goals) but I'd want to spin up my site's theme and point the underlying site to a 'testing' Micropub server so it can check that the site renders the theme and content types as expected when provided a set of MF2
#
jacky
it's like a whole theme testing thing I want to do (very specific to my setup) but the base bit about being able to test a page's HTML against a set of expectations is good enough for me
#
[aciccarello]
yeah, I'm hoping to get some jest.js tests looking at the parsed MF on my site eventually
#
jacky
def blog about it when you do!
#
[KevinMarks]
so, how do I make a <pre> full of code that doesn't wrap work on a narrow phone?
#
[aciccarello]
Like, are you wanting the code block to horizontal scroll?
#
Loqi
[Anthony Ciccarello] TIL you can emulate the as const type oberation in JS with a helper function. (source) /** * Identity function. Coerces string/number literals to value-as-type. * @template {string|number} T * @param {T} v * @return {T} */ function toConst(v) {...
#
[aciccarello]
And I noticed I misspelled operation. Good thing I have an edit button. 😆
#
[tantek]1
[KevinMarks] I think you're looking for white-space:pre-wrap https://developer.mozilla.org/en-US/docs/Web/CSS/white-space
#
[KevinMarks]
I did `pre {font-size:calc(min(2vw, 1rem));}`
[snarfed] joined the channel
#
[KevinMarks]
hm, so I have this 2 column design with text on the left and responses on the right. I'm now wondering about responses on the left, text in the middle and outbound links with previews on the right.
gRegor and kimberlyhirsh[d] joined the channel
#
[schmarty]
bringing jacky's example from #microformats and #indieweb-meta in: "something like http://lwa.tf/view?url=https://events.indieweb.org/2022/04/homebrew-website-club-pacific-Ohvjpl4IueP9 (not a real link) but when it's rendering a /rsvp, how to mark up the underlying h-event such that sub-RSVPs in the event would be picked up from a MF2 parser"
#
jacky
heh yeah
#
jacky
so it's mainly for marking up the attendees of a event (which is nudging me towards using `u-attendee h-cite` with a URL to a RSVP)
#
[schmarty]
it's not so important how they show up to an MF2 parser unless there's something using that parsed value. so like if a reader is going to show that content outside of e-content, then that logic in the reader seems like the first consuming case.
#
jacky
in the case of the reader, I do want to highlight RSVPs from authors/sites that fit a critiera (people/orgs I want to know more about, blocked, muted, etc). Having it as a `p-rsvp` kinda restricts me and makes more sense when publishing
#
[schmarty]
events.indieweb.org doesn't seem to have any mf2 around the RSVPs displayed there.
#
aaronpk
i believe i took the approach of no mf2 markup until it's clearly necessary for something
keyboardn00b[d] joined the channel
#
[schmarty]
i agree with that approach these days!
#
gRegor
My attendees just show as comments: `u-comment h-cite` and the `p-content` is something like "✔️ Attending
#
gRegor
I do group RSVPs separately from "Other Responses" though. https://gregorlove.com/2018/10/virtual-homebrew-website-club/#rsvps
#
[tantek]1
what are rsvps
#
Loqi
RSVPs are the responses displayed on an event indicating who is attending or not, might be, or interested in the event https://indieweb.org/RSVPs
#
[tantek]1
hmm... https://indieweb.org/rsvps#IndieWeb_Examples needs both cleanup and adding more examples (noted above)
#
gRegor
jacky, are you saying you want to discover those people/orgs from their RSVPs on events, or is it a known set of people already?
#
[tantek]1
I do think it would be a good / useful reader feature to show on an event "Here are the people you are following who have RSVPd to this event"
#
[tantek]1
the reader knows "people you are following". the event can publish "who have RSVPd to this event". so the reader, when displaying that event in a stream, and combine those two sources of information to show/prioritize showing who you are following that has RSVPd
#
[tantek]1
similarly readers could do this for "people you follow who have liked / reposted / commented on this post"
#
jacky
gRegor: the former!
#
[tantek]1
"who else is going to this event" is a common social question too, even if you're not following them
#
[tantek]1
A reader could present that information
#
jacky
[tantek]1: that's what I'm speccing out now too - with the known authors of a feed, one's set of /contacts and the like, "who are you 'familiar' with that's attending this event by a particular vector?" is one question I want to be able to answer
#
[tantek]1
FB event UI did this at some point IIRC, along with little "+" or "Add friend..." buttons next to the people you were not following
#
jacky
is also trying to bake in the relationship when showing these cues
#
jacky
[tantek]1: yup
tetov-irc and jacky joined the channel