#dev 2024-01-23

2024-01-23 UTC
btrem, tPoltergeist and geoffo joined the channel
#
btrem
I wish it were easier with 11ty to separate content from structure. It'd be nice to create a new github repo with the own-your-wordle code that I wrote, so others could use it.
#
btrem
Moving from #indieweb, re: https://btrem.com/feed/wordle
#
[aciccarello]
Seems like another space where I'd like to have content from personal sites listed or searchable by content type. Reviews, recipes, game scores.
#
[aciccarello]
But limited to the people I follow rather than a global search
#
btrem
I do post h-recipes, but in an unorthodox way. I'm actually thinking of moving my recipes to a sub-domain, and perhaps changing how it all works. Still trying to work out whether and how to change it.
#
btrem
Oh, and I do h-review for Lee Perry records, which I started after he passed away. https://btrem.com/2021/09-lee-perry-recommendations (but I haven't done any new ones in quite a while).
#
btrem
I just realized that I sound like I'm plugging my sites ("you should follow me"). Not my intention. Just chatting idly.
#
[aciccarello]
No worries
#
[aciccarello]
What issues have you had with 11ty?
#
[aciccarello]
I've thrown around ideas of moving my markdown files to a separate content repo and using git submodules to pull that into the build but seems like an unnecessary complication.
#
btrem
None with 11ty per se. (Well, lots with 11ty per se, but that's for another discussion!)
#
btrem
It's just hard to separate out one part of the site. Like the wordle stuff that I worked on this week. There are functions in .eleventy.js, and templates (what 11ty calls "layouts" but that's the wrong word and an example of 11ty problems per se!) in a directory. And style rules in a stylesheet that contains styles for the whole site. And directory data json and js files in yet another directory.
gRegor joined the channel
#
btrem
So how would I package that up so others could use it? I don't see how.
#
btrem
I've also thought of separating out .md files. It's crazy to have edits to a page (in a .md file) mixed in with changes to the site config. I even thought git submodules might be the way. But I'm an amateur. I know just enough git to get it all into a repo. Git submodules scares me.
#
btrem
I /think/ there's a way to
#
btrem
(sorry, hit the wrong key)
#
[tantek]
Ooh aside meta question, what presentational effect or semantic effect do you intend with the /word/ plain text btrem?
#
btrem
There's a way to create your own plugins, and I /think/ there's a way to pull a plugin from your own github account, but, as I said, I'm an amateur. Not sure how easy it would be to learn how that works.
#
btrem
[tantek] I assume you mean when I put a solidus slash / around a word, yeah? I mean italics, or to be more precise, emphasis in the html sense.
#
btrem
When I put a word between two solidus / marks, the word appears italicized in my chat client (Thunderbird). Same when I do it in email. Also in markdown, which I use for my website. I sort of thought it was a common convention. Does it not appear italicized in other clients?
#
btrem
I also use asterisks * around a word to denote strong emphasis, a la the <strong> element. And indeed, it shows as bold in the chat client just as <strong> does (by default) in a web page. And the * convention also works in email and markdown.
#
btrem
Maybe this discussion is better had in meta? Not that I mind, but others might find it tedious.
#
[tantek]
No no text markup and auto styling text is very on topic for #indieweb-dev
#
btrem
Ok, good. Maybe that's why Loqi hasn't scolded me yet. And here I was thinking she was getting soft on me. :-D
#
[tantek]
The /word/ shows up as literal "/"s around "word" in Slack (and I presume the archives)
#
[tantek]
I thought md did something wacky
#
btrem
Oh, yes, it shows up with literal / marks around it in my chat client, too. Same for email.
#
[tantek]
Oh yes markdown is wacky and treats asterisk * as italics
#
btrem
But markdown uses the / to change it from e.g. /word/ to <em>word</em>.
#
btrem
Oh, shoot, you're right. My bad.
gRegor joined the channel
#
btrem
One asterisk for <em>, two for <strong>.
#
[tantek]
You have to double asterisk ** to get bold which literally no one ever did in email before markdown. That was nonsense made up for markdown
#
btrem
I don't know the development of markdown. Maybe a solidus couldn't easily be parsed, so they went with asterisk instead? Dunno myself.
#
[tantek]
You have proven my hypothesis about "/" for italics or light emphasis and I will be citing what you said
#
[tantek]
In short, you're right, and markdown is wrong (about how to do italics in plain text)
#
btrem
Ha ha. Yeah, that's fair. You could even point out, when you cite it, that the person who mistakenly thought the / solidus means <em> in markdown was (and is) a regular markdown user.
#
[tantek]
btrem++ markdown--
#
Loqi
btrem has 2 karma in this channel over the last year (5 in all channels)
#
Loqi
markdown has -2 karma in this channel over the last year (-1 in all channels)
#
[tantek]
Btw in terms of chat clients, a single * adjacent to text (not spaces) starts a bold expression ended by another * adjacent to text which also makes sense to me (instead of the ** of md)
bret joined the channel
#
btrem
I notice there's a bit of hostility towards markdown. I imagine it's more than just the asterisk thing. As for myself, I find markdown much easier than e.g., wiki conventions. Maybe that's just what one is used to. And it's *much* better (see what I did there?) :-D than typing out html tags.
#
btrem
In fact, other than wiki and markdown, I don't know what other entry formats there are. I'd hate having to use a wyswig form. Don't want to keep grabbing my mouse to change elements.
tPoltergeist joined the channel
#
[tantek]
Wiki markup doesn't pretend to be plain text. Markdown does which is why the criticism
#
btrem
Ok, not sure I really understand what that means. I imagine it's not just the solidus vs asterisk thing?
#
[tantek]
Yeah markdown has many more problems than that. I've documented some (and possible fixes / replacements) on http://tantek.com/w/Markdown
bret, bret_, bret__, CRISPR, [KevinMarks] and [Al_Abut] joined the channel
#
[jeremycherfas]
I’d just add that the hostility to Markdown is not uniformly distributed. 😀
#
[tantek]
[jeremycherfas] do you think you'll make it to IWC Brighton 9-10 March?
tPoltergeist joined the channel
[jeremycherfas] joined the channel
#
[jeremycherfas]
I doubt it [tantek] I don’t want to fly and it is a lot of train for two days. Never say never, though.
Guest6, tPoltergeist and [tantek] joined the channel
#
[tantek]
Understandable. It is a lot of train travel though I wonder how cheap discount air tickets (EasyJet to Gatwick?) would be. And I understand not wanting to fly in general as well
[benatwork], geoffo, [Joe_Crawford]1, jjuran_ and [campegg] joined the channel
#
[Al_Abut]
[KevinMarks] good stuff in that list, thanks for sharing. I can use two of those tips right away to get rid of hacks I’ve been using for years - hacking up underlines with border padding and a non-breaking space to avoid word orphans.
[snarfed] joined the channel
#
[snarfed]
I'm back to thinking about de-duping protocol bridges. ie when a bridge first sees a user on a given network, how can it determine whether that user already has an account in the other destination network - whether bridged or otherwise - and if so, not bridge them?
#
[snarfed]
(related to but different than opt out, since that's separate per bridge)
#
[snarfed]
most protocols have native ways to indicate "also me/this" links - rel-me, FEP-fffd, NIP-39, etc - but I haven't yet come up with a way that's consistent across them. profile links are the closest I can think of, they might work. otherwise, maybe we just use each protocol's native way on its own?
#
Loqi
[preview] [snarfed] #800 De-dupe across different bridges
[aciccarello] and jonnybarnes joined the channel
#
[tantek]
[snarfed] also sounds parallel to rel-me identity consolidation
jacky joined the channel
#
jacky
using ntfy.sh doesn't seem like a bad idea
#
jacky
requires an unguessable secret (as the topic) and can be rotated often (secret hygenine)
#
[snarfed]
[tantek] yeah rel-me is the way for web sites, but other protocols generally have/need their own mechanisms
#
[tantek]
[snarfed] IMO then the answer is other protocols -> profile URLs -> previously solved problem with rel=me
#
[snarfed]
the problem is, there's no canonical profile URL for many other protocols, eg Bluesky or Nostr
#
[snarfed]
there are multiple clients, some web-based, but none are canonical. there aren't even well-defined (eg FYND) ways to discover _a_ public web profile URL, other than hard-coding them
#
[snarfed]
and you're also at the mercy of whichever web URL you load to include rel-me markup
#
[tantek]
hmm interesting
#
[snarfed]
each protocol's native mechanism is the alternative. not all of them have one, eg afaik ATProto doesn't, but others like Nostr do
#
[snarfed]
unrelated, has anyone looked at FedCM yet? looks like a "privacy preserving" alternative to OAuth? https://fedidcg.github.io/FedCM/ , https://github.com/fedidcg/FedCM/blob/main/explainer.md
#
Loqi
[preview] [Sam Goto] Federated Credential Management API
#
[tantek]
[snarfed] in short yes, have looked, and it's got ... issues. that's a longer conversation (maybe better in a sync medium)
#
jacky
I've been meaning to look into that for Sele
#
jacky
yeah, it looks like using what WebAuthn does for keys and slightly shapes it for accounts
#
jacky
very similar to GNOME Online Accounts, KDE PIM or whatever "system online accounts" equivalent one's system provides
lazcorp, lanodan and idealhack[d] joined the channel