#dev 2023-05-16

2023-05-16 UTC
[Tim_Nolte] joined the channel
#
[tantek]
would Micropub support fix this?
#
shreyanjain[m]
wait, does marsedit support micropub in any form?
KMF joined the channel
#
[tantek]
it does not. there was a session about it in 2017 and the conclusion was that Micropub would need additional features (or a complementary spec) for MarsEdit to consider supporting it: https://indieweb.org/MarsEdit#IndieWeb_Support
geoffo joined the channel
#
gRegor
oh, huh. I thought for some reason that had shipped. Must have misremembered.
#
gRegor
Might have been thinking of micro.blog's support, which was with XML-RPC
[manton] joined the channel
#
[manton]
I think Micropub has everything MarsEdit needs now. Just needs to be prioritized in one of his releases.
#
gRegor
Nice!
#
[tantek]
gRegor, from the toot you posted: "having a way to bulk-download a users' posts in their original format would be very helpful"
#
[tantek]
do we have a micropub extension for that yet?
#
[snarfed]
seems like there's been a long-lived tension in Micropub around whether rich read operations like this should come from HTML + mf2, eg fetching h-feeds and parsing next links etc, or from new Micropub operations
#
[snarfed]
seems like Micropub served CRUD use cases for individual new posts well, which drove some adoption, and now people want to use it for broader use cases like richer clients that surface historical data that wasn't originally created by those clients
epoch, [jeremycherfas] and gRegor joined the channel
#
[KevinMarks]
Looks like we need another pass at auth on bsky.link
angelo_ and [marksuth] joined the channel
#
[KevinMarks]
[capjamesg] OK, try that. I made it reauth if the refresh fails, and reset the timeout if we get a 401. There's probably a smarter logic along the lines Manton said using getSession, but this should get it out of the failed refresh without restarting.
#
IWDiscordRelay
<c​apjamesg#4492> Will review later.
#
[KevinMarks]
OK, but the sites down now, so at least restart it.
#
[KevinMarks]
also I found that resolveHandle doesn't need auth
#
IWDiscordRelay
<c​apjamesg#4492> Ah.
#
IWDiscordRelay
<c​apjamesg#4492> Deployed your fix.
#
[KevinMarks]
I left the logging on, so you may be able to see why if it fails again, unless there's some new reason
#
Loqi
[preview] There was an error fetching data from Bluesky. No thread or post
holiday_1 and bterry joined the channel
#
[KevinMarks]
Confirmed - the bluesky way of doing rich text is really annoying
#
[tantek]
they're not using HTML or Markdown?
#
[snarfed]
yeah it's start/end index into text content. oddly I thought AS1/2 do that too, and I've implemented it for them, but I can't find it now in their specs
#
[snarfed]
(and this is just for links and mentions, not formatting, bluesky doesn't have that in yet)
#
aaronpk
can't wait for the unicode/emoji char counting bugs to start popping up in clients
#
[snarfed]
that, they did a pretty deep dive on. I think they settled on counting rendered characters (apologies, I don't know the right term), not code points or bytes
#
[tantek]
glyphs?
#
[snarfed]
oh but their indices are UTF-8 bytes
#
[snarfed]
vs AS1/2 which I'm not sure specified 🤷
#
[tantek]
are they doing tco-like nonsense for the URLs as well or are they counting the full length of URLs?
#
[snarfed]
full length for visible text URLs, "free" for linked text
#
[tantek]
yeah that sounds like a disaster
#
[tantek]
"linked text"?!?
#
[tantek]
"indices are UTF-8 bytes" <-- there be dragons
#
[snarfed]
basically they count visible text, not formatting, including whether some visible text is linked
#
[snarfed]
do tell re dragons, I had to implement that recently
#
[snarfed]
on a related note, do you all remember the startIndex/length part of AS1/2? did I imagine it?!
#
[tantek]
could someone add a ref to Bluesky's "rich hypertext" docs to /Bluesky?
#
[tantek]
this does sound pretty nuts
#
aaronpk
what they settle on is irrelevant once a hundred different developers start trying to get it to work and all do slightly different things
#
[snarfed]
ah, sure. if you mean interop will be hard, yes
#
[snarfed]
not sure this is documented yet outside of code
#
[tantek]
precisely aaronpk. "indices are UTF-8 bytes" is a developer footgun
#
[snarfed]
right, can you explain or cite? I'd love to know more
#
[tantek]
requires having to know how your specific language handles Unicode strings
#
[snarfed]
ohhhh sure. ok
#
[tantek]
which is usually below the level of abstraction that developers work with, for good reason
#
[tantek]
strings should "just" work, using built-in language string level operations
#
[snarfed]
sure! but I think a bigger fraction of developers are now at least aware of character encodings, and some of the basics, vs 10-20y ago
#
[snarfed]
in general though yes this stuff isn't easy
#
[tantek]
awareness is less of a problem sure, getting it right with all the quirks of Unicode, good luck. great way to introduce fragility into a format
#
[tantek]
offsets-- in general for any kind of string API
#
Loqi
offsets has -1 karma over the last year
#
[snarfed]
I know a decent amount about this stuff, and amusingly I had a much harder time counting glyphs than bytes
#
[snarfed]
but yeah agreed, offsets--
#
Loqi
offsets has -2 karma over the last year
#
petermolnar
"indices are UTF-8 bytes" 
#
[snarfed]
so startIndex/length in AS1/2, did they not exist? did I imagine them? I haven't found them in any of http://www.w3.org/TR/activitystreams-core/ http://www.w3.org/TR/activitystreams-vocabulary/ https://activitystrea.ms/specs/json/1.0/ https://activitystrea.ms/specs/json/schema/activity-schema.html . were they in some tag extension that I've forgotten about?
#
sknebel
I mean counting glyphs is basically impossible unless you know *a lot* about the output chain
#
sknebel
including font selection and rendering
#
Loqi
[preview] [James M Snell] Activity Streams 2.0
#
sknebel
byte-wise is IMHO overall the safest option if you insist on doing index-based
#
IWDiscordRelay
<c​apjamesg#4492> Did Twitter have an open source algorithm on how they counted characters?
#
sknebel
yes, they had documented that
#
IWDiscordRelay
<c​apjamesg#4492> I remember reading something about it.
#
IWDiscordRelay
<c​apjamesg#4492> What is a character count?
#
IWDiscordRelay
<c​apjamesg#4492> That’s what I read.
#
sknebel
(shows some of the pitfalls, e..g defining the emoji-exception specifically as "emoji supported by our emoji library" == implies knowledge about rendering)
#
IWDiscordRelay
<c​apjamesg#4492> Yes indeed.
#
bkil
[snarfed]: As AP (and Fediverse protocols preceding it) are already working through HTML by default, adding an extension that decorated sections of the marked up HTML using byte offsets would sound a bit weird.
#
bkil
I think some widely used systems that were doing that were Java SWT StyledText, Apple NSTextView and some of Microsoft's widget toolkit, not sure which one https://github.com/bkil/wiki/blob/master/en/widget-toolkit-html-rich-text.md#user-content-java-swt-styledtext
#
[KevinMarks]
Also that can change with unicode releases.
#
[KevinMarks]
Anyway the link code I wrote this morning fails when there is an emoji before the link
#
[KevinMarks]
Also if you chop up utf8 at byte boundaries weird things happen
#
bkil
[KevinMarks]: Yeah, the Matterbridge bot transforms the whole message into garbage if the end is chopped off due to IRC limits in the middle of a multi-byte sequence.
[capjamesg], [Murray] and angelo joined the channel
#
bkil
[tantek]: So staying constructive, what syntax would you use to insert a multi-word phrase to a URL?
#
bkil
Adding parenthesis itself is extra cruft if you interpret it strictly. Shouldn't we then just anchor-link a URL to the preceding word everywhere where it is not on its own sentence or line?
#
bkil
"almost the same" vs. dogfooding - people had already been using Markdown for quite some time now and code talks.
#
bkil
If you want readability, adding ^ to footnotes is already cruft. Why not convert every inline URL to footnotes automatically when rendering? That would be a 0-syntax alternative.
#
[tantek]
yes it has appeal among a certain nerd set, but so does HTML
#
[tantek]
there is a semantic and presentational difference between inline links and footnotes
#
[tantek]
and not all footnotes are (just) links
#
[tantek]
this is pretty self-evident when you have actually published a bunch of content with both
#
bkil
self-evident vs. "just"
#
bkil
What is just?
#
Loqi
It looks like we don't have a page for "just" yet. Would you like to create it? (Or just say "just is ____", a sentence describing the term)
#
[tantek]
no, empricial experience vs theoretical "why not"
#
[tantek]
empirical*
#
[tantek]
more on the footnotes syntax explorations is reasonably well documented in /footnotes
#
[tantek]
re: "anchor-link a URL to the preceding word everywhere where it is not on its own sentence or line" — I've considered it, but with a bit of existing content analysis (nothing comprehensive), there were enough false positives that I've rejected this alternative myself
#
bkil
It is "self-evident" to me that a huge majority of sites worldwide not created using WYSIWYG are statically rendered or edited using mediawiki, CommonMark or RST, not some syntax that does not exist yet.
#
bkil
I also see that CMS the above markup types are way more popular than people editing HTML by hand.
#
[tantek]
I don't think that's what self-evident means, because you're applying analysis.
#
[tantek]
there's plenty of prior art to consider yes
#
[tantek]
indeed MediaWiki syntax is less crappy than Markdown IMO
#
[tantek]
more usable
#
[tantek]
what is RST?
#
Loqi
It looks like we don't have a page for "RST" yet. Would you like to create it? (Or just say "RST is ____", a sentence describing the term)
#
bkil
I'd love to look into your prior analysis on autolinking if you have it published anywhere.
#
bkil
RST is reStructuredText
#
[tantek]
yup, my own notes posts from 2010-present
#
[tantek]
as well as analysis of what Twitter tended to auto-link (differently over the years)
#
[tantek]
some of that was captured in my own posts analyzing it, some on /Twitter I believe
#
bkil
It took me way more time to learn MediaWiki than CommonMark and I still mistype it more often. {} {{}} [] [:] [[]] [|] [[|]] ...
#
[tantek]
hah that's real
#
bkil
[tantek]: Feel free to send me a URL or two if you find the analysis.
#
[tantek]
not sure what you mean bkil, already answered that question above
#
[tantek]
I typically blog (post notes) about bits of analysis as I come across them, and my own plain text notes (in plain text form) have been useful as source material (when I'm only thinking about writing, not structuring anything)
#
bkil
I would like to see your false positive and run it through my parser to see whether I made the same mistakes.
#
[tantek]
Feel free to go through those
#
[tantek]
you may also be interested in the methodology of https://indieweb.org/text-first_design
#
[tantek]
RST has some interesting ideas in it. Its inline markup is pretty poor, but its block-level markup / processing has some promise.
#
[tantek]
What CMS uses RST?
#
aaronpk
i've seen it used in generators for docs websites and print books
#
[tantek]
What if any IndieWeb projects use RST? The page on the wiki is a pretty minimal stub with no IndieWeb relevance (yet)
#
bkil
Python Sphinx, Trac, GitHub, BitBucket, Distributed Proofreaders, Linux kernel, CMake is mentioned here https://en.wikipedia.org/wiki/ReStructuredText#Applications
#
[tantek]
if it doesn't have any IndieWeb-specific relevance that anyone knows of, then we should delete and merely link to the Wikipedia article instead
#
bkil
I think CommonMark did intend to codify email conventions from the 80s, but could you tell me why underscore was used for /italic/ text instead of underlining?
#
[tantek]
does anyone use RST to write content for their personal site for example?
#
[tantek]
no I cannot and that is one of my complaints against Markdown/CommonMark
#
[tantek]
Markdown started with a good principle:
#
[tantek]
"The overriding design goal for Markdown’s formatting syntax is to make it as readable as possible. The idea is that a Markdown-formatted document should be publishable as-is, as plain text, without looking like it’s been marked up with tags or formatting instructions."
#
[tantek]
and then promptly violated it with link syntax, misuse of underscore for italic text and more
#
[tantek]
plain text email is a reasonable source of implied structure from lightweight punctuation
#
bkil
'''Definitely'''
#
[tantek]
I'm actually considering splitting the inline vs block level implied structure from punctuation into two separate specs/libaries because so many different "lightweight markup" (so to speak) languages have better/worse approaches to one or the other
#
[tantek]
my guess is that many would appreciate using only an inline-level auto-linker (and auto-structuring) library while "leaving alone" any of the block level spacing / linebreak semantics
#
[tantek]
One challenge I have found is too many false positives of mistructuring the use of an asterisk as a mini-inline note, i.e. trailing* like that and then eventually on a line by itself (*trailing meaning immediately adjacent to the end of a word)
#
[tantek]
interesting, CMoS 6.110 discourages use of angle brackets for delineating URLs.
geoffo joined the channel
#
[tantek]
IMO anyone designing a punctuation based plain text format SHOULD read CMoS
#
capjamesg
Oooooh!
#
capjamesg
Have you read it [tantek]? (Serious question; I'm tryin gto evaluate whether, if I buy it, I will likely read it)
#
capjamesg
Or would you consider it more of a reference material?
#
[tantek]
Yes and yes
#
[tantek]
It's very long so it's definitely a skim-read to get the general idea of the kind of specific guidance in it, and a feel for where to look things up later when you need to
#
[tantek]
It's definitely reference material
#
capjamesg
Ordered.
[chrisbergr] joined the channel
#
[KevinMarks]
Link offsets are bad, continued
geoffo joined the channel
#
[tantek]
looooool
#
[tantek]
^ my CASSIS auto_link function handles that source text just fine, in JS & PHP :)
[Caleb_Hearth] joined the channel
#
bkil
So, I don't think that CommonMark is the only way to go, neither do I think it is the way to go, but I've yet to see a competitor I could adapt.
#
capjamesg
[tantek] I can't find it but I feel I have expressed my like for markdown in chat before.
#
capjamesg
Markdown, in addition to your footnote trick, works really well for writing for me.
#
bkil
I'd be happy to adopt anything for my needs. I would need an extension for gemini: something that is easy to reason about and easy to part, but with the basics also included (not the bare essentials that they have targeted)
#
bkil
easy to parse
#
Loqi
[preview] [capjamesg[d]] I think well in markdown 😂
#
[tantek]
again, if you're prioritizing "easy to [machine] parse" you have the wrong prioritization
#
bkil
Both
#
capjamesg
The syntax is minimal enough for me while retaining readability during drafting.
#
bkil
Easy to parse by both machine and humans. The CommonMark dialects handle subtle formatting with newlines differently and that confuses me as a reader as well.
#
[tantek]
"during drafting" wasn't the point. did you read [chrisbergr]'s statement that I cited?
#
[tantek]
yeah I'm hesitant to deal with the newline formatting inconsistencies
#
[tantek]
that screws up a bunch of scenarios
#
capjamesg
I know. I'm making a separate point.
#
[tantek]
CommonMark-- for inconsistent (hard to remember and hard to predict) formatting with newlines
#
Loqi
CommonMark has -1 karma over the last year
#
capjamesg
That is one reservation I have: I have encountered a lot of friction with bold vs. italics.
#
Loqi
[preview] [capjamesg] I just realised I have been marking up questions in my interviews as bold text in markdown.
#
bkil
When I had to come up with a lightweight markup on my own way back, it was very similar to markdown (it was not widely know around then) At the same time, I would have never came up with the syntax oddness that we call mediawiki.
#
[tantek]
yeah Markdown really screwed up bold & italics
#
capjamesg
I now use _ text _
#
capjamesg
for italics\
#
bkil
Looking the wiki chart posted earlier about alternatives, I can share a definite things about how things should not look like. At the same time, I could accept quite a few alternatives that I could happily adapt.
#
capjamesg
This was an interesting day in markdown: https://chat.indieweb.org/dev/2023-04-16
#
[tantek]
I think it's a mistake to start with Markdown. This is why I started with auto-linking. This is the comparison people are missing.
#
[tantek]
auto-linking is implemented and used by FAR more implementations and people than Markdown or any other plaintext-corruption format
#
bkil
Yeah, I just added autolinking to my wiby chat client - it took just a few lines of code: https://bkil.gitlab.io/static-wonders.js/userjs/ghostchat-wiby.html
#
capjamesg
More in markdown discussions ^
#
capjamesg
Wait, I lost one of the URLs.
#
capjamesg
:facepalm:
[schmarty] joined the channel
#
[schmarty]
enh, people are going to use whatever formats they feel like learning and sticking with. there are loads of reasons things "stick". generations of bbcode, incompatible wiki syntaxes, flavors of TeX and more have come and gone and will continue to do so.
#
[tantek]
capjamesg, which points in particular did you think were interesting? I feel all I got out of that discussion was encouragment to create a fragment permalink: https://tantek.com/w/Markdown#PrimeDirective
#
[schmarty]
i write a lot of markdown, mostly in whatever subset seems to work most places. i have to look up mediawiki syntax like 10x more often. 🙃
#
capjamesg
The indents.
#
[tantek]
[schmarty] yes, lots of these things come & go. My point is you don't see people having to put extra effort into "learning" auto-linking. It's something that "just works" and people deal with just fine without having to think about it
#
[tantek]
same thing with using @- or #- to mean something
holiday_1 joined the channel
#
capjamesg
Yeah. I have had more trouble with the MediaWiki syntax (although recency bias may play in there - introduction to mediawiki ~3 years ago vs. markdown way further back)
#
[schmarty]
tantek: as a counterpoint i struggle with my very own autolinking rules on my site all the time, both remembering (how) to use them and with making sure they don't break :zany_face:
#
bkil
I think BBCode was a hack for people too afraid to parse and sanitize HTML properly, but it's not very user friendly. Neither would I recommend LaTeX for normal people, although I did take all of my notes in it on my pocket computer on lectures!
#
[tantek]
haha developer problems 😂
#
capjamesg
I apologise for being a bit less constructive than usual; it's been a rather stressful evening.
#
bkil
Oh, autolinking just works is also a myth. Didn't I share the story about the Matrix client that "fixed" autolinking in a way that made Matrix room link break (because they include two hashmarks)? And the main developer was vehemently opposed to fixing it, despite me pointing to RFC text that clearly pointed at how it should be working.
#
[tantek]
the people that are good at LaTeX are impressive and I think it's one of those things that once you get good at it, it affects how you think when writing
#
[tantek]
Matrix using hashmarks in URLs is a bug in Matrix
#
bkil
Well, yes and no.
#
[tantek]
[schmarty] clearly we need to join forces on an auto-linking test suite 🙂
#
bkil
Let me find the references for you, back in 5.
#
capjamesg
What about groff?
#
[tantek]
look, I thought we were passed all the hashbang nonsense
#
[tantek]
hashbang--
#
Loqi
hashbang has -1 karma over the last year
#
bkil
TL;DR, if you don't break on spaces, you will have a bunch of ugly corner cases.
#
capjamesg
Hashbang?
#
capjamesg
(in this context?)
#
[tantek]
capjamesg be grateful you missed that era 😂
#
capjamesg
Brian Kernighan wrote "Unix: A History and a Memoir" in groff I think.
#
capjamesg
He mentioned it in a Computerphile video.
#
capjamesg
I agree with your point re: LaTeX.
#
Loqi
[preview] [capjamesg] h-resume-to-man-page: Convert a h-resume document to a Linux man page.
#
[chrisbergr]
Because I know html on an advanced level, I would always choose textile for my plaintext markup. https://textile-lang.com/
#
[schmarty]
tantek: whether footnote or bracket or dangling unicode superscript - i would argue that all linking (including autolinking) is a syntax and it must be learned and remembered
#
[tantek]
[schmarty] the key to designing good auto-linking is to avoid surprises, especially anxiety-inducing surprises. As long as you can feel "safe" typing / pasting in whatever plain text you want, and if auto-linking(or embedding) happens the feeling should always be neutral to positive.
#
[tantek]
The goal should be not having to feel like you have to learn or remember anything in particular. The plaintext forms should "look fine" as-is, and if they're enhanced with auto-linking, their fundamental meaning should stay the same and it should "feel like" a presentational enhancement.
#
[schmarty]
fine goals to choose and aim for!
#
[tantek]
these are goals that Markdown (and many/most other "plain" text "formats") very much abandoned
#
[chrisbergr]
tantek: As far as I get it you use plaintext for your website with autolinking, auto linebreak and so on. But without markdown? How do you do bold, italic, images on your site?
#
[schmarty]
in other news, i think i have just about finished porting the 🕸 from nodejs to PHP ??? a couple more tweaks to the gardening and then i'll need to decide how and when i am gonna roll it over.
#
bkil
https://datatracker.ietf.org/doc/html/rfc3986#section-2.2 "If a reserved character is found in a URI component and no delimiting role is known for that character, then it must be interpreted as representing the data octet corresponding to that character's encoding in US-ASCII."
#
bkil
https://en.wikipedia.org/wiki/Uniform_Resource_Identifier#Syntax - So basically yes, your URI can very well contain multiple hashmarks and a bunch of stuff that regularly breaks autolinkers that are too "smart"
#
[tantek]
[chrisbergr] I don't do bold currently. I am currently considering `*bold*` because that's what works in various chat programs already. Not the stupid double ** thing from Markdown which literally no one did in email. That was made up from whole cloth.
#
bkil
[chrisbergr]: In my MVP social networking app, along with supporting HTML and a CommonMark subset, I have enhanced autolinking to also offer inlining images, videos and iframe'ing other rich external content based on the file extension, path, subdomain and domain. It could be further enhanced via OEmbed as many Fediverse servers are doing it. So basically, an author just posts a URL, while the reader will see an image (video player, shader animation, etc.) if
#
bkil
they have it enabled in their preferences.
#
capjamesg
I like *text* for bold.
#
capjamesg
And _ text _ for italics.
#
capjamesg
I think that's how Discord does it.
#
[tantek]
bkil, you may want to instead reference the URL Living Standard https://url.spec.whatwg.org/
#
capjamesg
And Slack.
#
[tantek]
capjamesg, exactly, that's how Slack and Discord do it
#
bkil
capjamesg: Yeah, something to think about. I sometimes feel that doubling characters for your markup is a bit of "cheating" and adds to the line noise.
#
capjamesg
I don't understand the ** **
#
bkil
See the charts of others who support your ways. I'm a bit disappointed that matrix chose to double stars for boldness https://en.wikipedia.org/wiki/Lightweight_markup_language#Comparison_of_lightweight_markup_language_syntax
#
bkil
(matrix as in the most commonly used matrix clients, of course the protocol itself uses HTML)
#
[chrisbergr]
bkil: That means if my image url does not contain the file extension for an image it would be displayed as an url, right?
#
capjamesg
Oooh, nice chart!
#
Loqi
bkil has 3 karma in this channel over the last year (9 in all channels)
#
[chrisbergr]
tantek: * makes so much more sense than **
#
[tantek]
[chrisbergr] yes. those two replacements are top of my replacements for Markdown table: https://tantek.pbworks.com/w/Markdown#BeforeandAfter
#
bkil
[chrisbergr]: Yes. Also assuming the domain or subdomain is not a well known one (can be edited by the user on a whitelist basis) I'm still playing with the UX concept of adding a "preview" button that would successively try to OEmbed/img/video/audio/iframe whatever the link points to in sequence for a few more round trips.
#
bkil
Could you perhaps share the use case where this would be a problem for you?
#
[chrisbergr]
You are using cassis for all of this, right? I just remember that I used it also many years ago before I dropped my self coded website for my wordpress site.
#
bkil
Even chat-uploads.iw adds an extension thankfully.
#
[tantek]
images are further down in that table. in summary, the way people do in chat / txt. if you drop a URL to an image resource, it gets embedded. no need for any bang-junk
#
[tantek]
Yes the portions I use are implemented in CASSIS.
#
bkil
Yes, and it works just as I metnioned: it tries to generate a preview for it (server side) via OEmbed or fetching the resource directly.
#
bkil
At the same time, if you are working on a backend-optional app, the image is huge and the user would like to conserve bandwidth, this might not always be the best way to go.
#
bkil
If I'm producing rich content (i.e., a message akin to "HTML-email"), I sometimes desire to embed items into specific positions resized in specific ways. CSS selectors also allow you to load the version of the image that will result in the minimal amount of download as well via media queries.
#
[chrisbergr]
bkil atm I don't have a valid use case. I had issues with this using my own image uploader so I changed the urls providing a file extension https://images.cho.bz/uploads/1681858936.jpg?w=1280&fit=crop&q=100&sharp=10&dpr=1&s=f479fa32969398f8c253e96fee41ad84
#
[tantek]
[chrisbergr] the "controversial" bit of "syntax" that I'm considering is [URL] to mean "always embed this URL, like in an iframe, or special case img tag for URLs with image format file extensions"
#
bkil
Another thing about markup. I always found the concept of highlighting select parts of words fascinating, but the more content I generate the more I realize how useless this is. I.e., I can 100% live without such a feature and it makes the parser way simpler.
#
[tantek]
since the brackets [ ... ] sort of look like a frame around a picture
#
[snarfed]
[tantek] fwiw Wordpress uses that syntax for embeds too, it calls them "shortcodes"
#
[tantek]
wait what
#
[snarfed]
I think
#
[chrisbergr]
tantek yes, visually this makes sense and it is easy to remember
#
[chrisbergr]
[snarfed] [] is for shortcodes
#
[tantek]
[snarfed] no WordPress uses [ ] instead of HTML < >
#
[snarfed]
oh, not via URL though, true
#
[tantek]
that's not what I mean
#
[snarfed]
https://codex.wordpress.org/Shortcode . also has a params syntax, but that's optional
#
[snarfed]
ah ok nm then
#
[tantek]
I mean literally [https://example.com] means put an iframe here to http://example.com
#
[tantek]
that's it
#
[tantek]
this is not plain text: [embed width="123" height="456"]...[/embed] (from https://codex.wordpress.org/Embed_Shortcode)
#
[tantek]
if you're going to do that, just use HTML <img width height src /> — it's less typing
#
bkil
Just looks like forked BBCode.
#
capjamesg
What is bbcode?
#
Loqi
It looks like we don't have a page for "bbcode" yet. Would you like to create it? (Or just say "bbcode is ____", a sentence describing the term)
#
[tantek]
yea I was going to say. BBCode uses [ ] while SGML/HTML/XML uses < >
#
[tantek]
does anyone use BBCode on their personal site content?
#
[chrisbergr]
I'm thinking about implementing [spoiler]...[/spoiler] on my site 🙂
#
[tantek]
like visible to readers or only to you as author of the content?
#
[chrisbergr]
And also I want some kind of trigger warning overly thingy... I also thought about some bbcode syntax for this.
#
capjamesg
Oh no haha.
#
capjamesg
(re: bbcode syntax)
#
[chrisbergr]
tantek: bbcode visible for me and for the readers this should be some expandable box
gRegor joined the channel
#
[tantek]
[chrisbergr] this makes sense. Want to stub a page for /BBCode with your brainstorming? An IndieWeb desire / intent to implement is very much relevant to IndieWeb and worth capturing (even if you don't get to it for years 🙂 )
geoffo joined the channel
#
[tantek]
what is a spoiler
#
Loqi
spoiler post is a post that contains information that will spoil something such as a movie, in the IndieWeb this would be a post that communicated to a parser/reader that the content of the post might spoil info about a new book, tv show, movie, etc https://indieweb.org/spoiler
#
[tantek]
what is a content warning
#
Loqi
content warning is a feature of a post create UI where an author can hide some or all of the primary content of a post due to concerns about negative impacts of the content upon viewers, and typically provide a text warning (eponymously named) with the nature of the concerns, which is initially displayed instead of the post content https://indieweb.org/content_warning
#
[chrisbergr]
Yes sure, if you can quickly create this page with the loqi promt I can edit the page with my plans/ideas 🙂
#
[tantek]
Just missed the window but I'll click the Loqi link after capjamesg asked
#
bkil
[chrisbergr]: See <details> and <summary> https://developer.mozilla.org/en-US/docs/Web/HTML/Element/details
#
[chrisbergr]
I guess I need to learn how this wiki prompts work 🙂
#
[chrisbergr]
bkil oh thanks, that would be the perfect output for my implementation 🙂
#
[tantek]
here you go [chrisbergr]: https://indieweb.org/BBCode 🙂
#
IWDiscordRelay
<c​apjamesg#4492> ++
#
bkil
The markdown syntax leaves a lot to be desired and not many systems support it - StackExchange, Reddit ( >! ) https://meta.stackexchange.com/questions/1191/add-markdown-support-for-hidden-until-you-click-text-aka-spoilers/71396#71396 https://old.reddit.com/wiki/markdown
#
bkil
[tantek] [chrisbergr] : By the way, did you know Markdown already supports references via `whatever[1]` and then adding `[1]: http://example.com` at the end? It usually doesn't render as a subscript by default, but it's a trivial tweak to add. I also find it readable due to it occurring in free text a lot in scientific papers.
#
bkil
superscript
#
bkil
I've never encountered the alternative of whatever^1 in the rendered and printed form yet (thankfully), so that would probably surprise me.
#
[tantek]
the use of ^n to indicate exponent n or superscript n in single-line ASCII7 is quite old
#
[tantek]
I don't know if anyone else repurposed it previously for footnote superscript numbers, but it seemed like a fairly obvious thing to do
#
bkil
I didn't get a response at that time. But I'm sure you've also shared your idea about the caret as well. Yes, and Haskell uses it for exponentiation.
#
[tantek]
lack of response is a good indicator of lack of interest, no need to bring it up again unless you have new information like you've implemented it or something
#
bkil
I usually prefer the form whatever^{1} for clarity, though, but that actually results in worse results according to your taste.
#
[tantek]
the discussion of ^n for footnotes was a consequence of folks reading my posts actually using them, with real world permalink uses to reference and analyze in context
#
[tantek]
in general folks in the IndieWeb tend to prefer spending time on practical discussions with real world examples with personal permalinks more than ideas / preferences out of context or without any intention to implement.
#
[tantek]
"your taste" is narrow framing, see /footnote for broader documented feedback/suggestions
#
bkil
"Your" interpreted as in "IndieWeb" isn't that narrow in my opinion.
#
[tantek]
that sounds meta/abstract enough to be unclear what the point is, certainly more for #indieweb-meta than #indieweb-dev
#
bkil
I don't have a strong opinion on this matter right now as nobody I interact with uses are prefers footnotes.
#
bkil
I'd like to add some more input then to the question. The Opera keyboard shows `[a-z./ ]`, backspace, cancel and enter by default, `[0-9()+*=&_'"@?!;:,%#-]` on the second page and can finally access`[][$<>{}^\\~|]` through another touch. That would make using / for /emphasis/ oddly easy. How does the layout of your on screen keyboard look like?
#
[KevinMarks]
As noted in https://indieweb.org/fragmention#rejected_double-hash_URL_style the URL spec doesn't allow # in fragments, expecting it to be escaped. Empirically we found some apps ended a URL at the 2nd # when inlining it.
#
bkil
A Blackberry seems to offer `[a-z$ ]` on touch, `[0-9()+*_'"@?!;:,#./-]` with `alt` and others through the `sym` button https://upload.wikimedia.org/wikipedia/commons/0/09/BlackBerryKey2LE3.png That would make utilizing $ in markup pretty convenient as well (along with coding in Bash, Perl or PHP). MathJax and Katex comes to mind.
#
bkil
Are any of you blogging about math topics or over a mobile phone?