LoqiIt looks like we don't have a page for "Fragmentions and Verify Me" yet. Would you like to create it? (Or just say "Fragmentions and Verify Me is ____", a sentence describing the term)
LoqiA fragmention is an extension to URL syntax that links and cites a phrase within a document by using a URL fragment consisting of the phrase itself, including whitespace https://indieweb.org/fragmention
Loqiverify-me is a browser extension that checks if rel-me links on the current page link back, and shows badges if they do https://indieweb.org/verify-me
[KevinMarks]What does the new manifest make harder ? I know part of the motivation was limiting ad blockers. Both fragmention and verify-me do dom parsing and updating, and verify-me needs to fetch arbitrary urls too
btrem(A pedant says: the mime type for html -- the most common language on the web -- is text/html, not text/plain.) So do you want to display either a web page (text/html) or json for any given url?
btremNote that conneg by content type is very frowned upon here. I don't know how to check its karma score, but it's always negative double digits. I could downgrade it again, and then Loqi would give us the updated tally, but I don't hate it as much as others here.
btremI will point out that conneg by content type was fundamentally broken in the olden days of MSIE 5. Intentionally, AFAICT, by MS. So there's that. MSIE is not longer relevant, but there are other problems, in particular with caching, that will create real hurdles for you.
[tantek]I don't think we did anything "special" with conneg in IE5/Mac for example. We did support it well enough to let webdevs use PNG 1.0 with alpha as an upgrade (higher quality) from GIFa without proper alpha.
btrem[tantek]: ah sorry, should have specified. I was referring to MSIE 5/Win. The Mac code base was completely different. (Which you obviously know, but others may not.)
btremAnd what do you mean by a text browser? I don't know of any browsers that prefer text/plain over text/html, which will present a real hurdle to do conneg the way you want.
btremAs for curl, yeah, I suppose plain text would be easier to read in terminal output compared to text enclosed in html tags. But how many people read the web by issuing curl commands in a terminal, except to test something out?
btremThe most obvious difference is that using Lynx, I can follow links (<a href="foo">foo</a>) by navigating to the link and selecting it. And links are pretty fundamental to the web. With plain text and curl, I'd have to issue a new curl command, and copy-and-paste any links I found.
[Murray]Though you are wrapping everything in a `<pre>` element? Obviously, it's your site and you should make it how you want it to be, but I'll add that this makes it fundamentally pretty inaccessible. So you _might_ be aiding CURL, but you're cutting off anyone using non-visual browsing methods 😅
[tantek]Eg I can type example dot com like: http://example.com and Slack turns that into a hyperlink. I have no idea (don't care) of the plumbing it used to do so
[Murray]I feel like that's fine in an editor context, but Twitter has been a very good example of why you shouldn't do it automatically 😄 Who was the politician that kept missing spaces after periods, and people would buy up the resultant domains and forward them to NSFL content or opposing political views etc?
btrem[tantek]: Not so sure I like that idea. That's taking a page advertised as text/plain and deciding it's something else. Slack is not, AFAIK, capable of html, so adding autolinking might make sense there. (Don't use Slack, so I don't know.) But a web browser is. If someone wants autolinking, then use what already works, text/html.
[Murray]I also (personally) have a long-standing battle with this in a tool I use daily, WorkFlowy, where it would regularly hyperlink things that I don't want hyperlinked. I'm literally looking at a note right know on http://webmentio.io, and it's been hyperlinked automatically, which I cannot override and do not want
[Murray]what you're proposing removes consent from the equation, and relies on a machine to be able to do it accurately to prevent abuse vectors, which I do not think is (currently) possible
btrem[tantek]: I cannot confirm that View Source turns plain text into links. Does /not/ work in Firefox, nor Chromium, on Ubuntu. Or did you mean it allows users to follow links in href attributes?
gRegorFYI, the Fragmentions extension in Chrome got turned off automatically for me. Chrome says "Off • This extension hasn't published privacy practices, such as how it collects and uses data
LoqiA fragmention is an extension to URL syntax that links and cites a phrase within a document by using a URL fragment consisting of the phrase itself, including whitespace https://indieweb.org/fragmentions
btremIf you'll excuse the pun, autolinking plain text in a browser smells like content-sniffing. Yes, it's different, but the principle is similar: the browser second-guessing the author. Content-sniffing--
[Murray]if we're talking about autolinking an href attribute, that also makes sense to me, as authorial intent is clear and you can be reasonably sure that the link is, in fact, a link.
btremI got that. But I put a plain text link in a page, then viewed source. I did not turn it into a link, any more than it did on the regular view. And as I said, href= does not constitute plain text in my book.
LoqiURL design is the practice of deliberately designing URLs, in particular, permalinks, typically for a better UX for everyone who creates, reads, and shares content https://indieweb.org/URL_design
[lazcorp]My site isn't case sensitive on URLs for pages (which go through my routing system), but it is case-sensitive for file names of files hosted on there
btremcapjamesg[d]: I don't think there's any way around the problem. It's also a problem for domain names. So imagine a domain name for a river, say Concord River. It might be concordriver.example.org or concord-river.example.org. Either of those might be read as "concord river dot example dot org.
[tantek]only valid use-case I have found for lowercase camelCase in prose is in multiword hashtags because it can increase the chance of them being read out loud more accurately by text to speech or other accessibility tools
[tantek]capjamesg[d] re: two-or-more-word-long URLs, assuming you mean in the path, it used to be convention to hyphenate for better readability and search indexing, and now it's kind of a toss-up (personal pref) between - and _ delimiters
[lazcorp]Hmm, I guess that static sites are nearly always going to have case sensitive URLs, as every page is an actual HTML file (as opposed to the request being routed from a front controller through to a method in a specific controller), and 99.9% are going to be hosted on Linux
btremI've never actually done that. I sort of think it's wrong to server the same content at /foo and /FOO. If I never publish a link to /FOO, I don't see the need.
btremAlthough I read the page, and the links-in-bio, I'm not sure what they do. It looks like their contact pages? But without the customary form? But not always?
btremI'm having a face palm moment. (Actually, it was a couple of days ago.) I post Wordle results on my web site, and in the interest of accessibility, I replace unicode tiles with an svg approximation of them.
btremBut it occurred to me, why not just wrap the unicode tile in a `<span>` and add an `aria-label` attribute with the same text? In which case, why did I go through all the trouble of hand-coding svg to look like tiles? Gah!
btremReminds me of the time I wrote a function in php to output database results as json. Took me several hours over a couple of days if memory serves. Then a day or two after finishing it, I discovered json_encode. :-/
btremHmm, does anyone use web text to speech software? Maybe just to test things? I'd like to see whether `<svg><title>foo</tltle><!-- tile shape --></svg>` is better than `<span arial-label=foo><!-- tile --></span>`.
[Murray]I regularly test with screenreaders and other assistive tech, and `<title>` is poorly supported. ARIA is a generally better option, but know that ARIA labels on `<span>` elements is actually invalid, unless you also provide the relevant role (in this case I'm guessing "image")
[Murray]I'd say, if you're wanting to make something like a Wordle grid accessible, treating the whole grid as a single image and giving it all a descriptive label (best case: visible caption; second best: invisible caption, without removing it from the DOM) would be best
[Murray]something like "Wordle grid with four attempts. The first shows two correct letters in positions 1 and 4; second row has three correct in 1, 2, and 4..." etc. (the fewer words the better; I'm bad at that part 😄 )
NaomiAmethyst7, NaomiAmethyst and btrem joined the channel
[Paul_Lieberman]Hello #indieweb Dev . I've got the Drupal modules installed and I'm trying to setup authentication to use my website. If I try logging in at http://indielogin.com it redirects to my site just fine. I click Authorize and it goes back, but gives an error: "Missing or invalid parameters". Any ideas how to go about troubleshooting this? Thanks.
[KevinMarks] that's annoying - chrome finally turned off the old manifest format
Loqi It looks like we don't have a page for "Fragmentions and Verify Me" yet. Would you like to create it? (Or just say "Fragmentions and Verify Me is ____", a sentence describing the term)
capjamesg[d] What is fragmention?
[aciccarello] [KevinMarks] , I can hook you up with a beta version of Omnibear using manifest v3
[Murray] Though you are wrapping everything in a `<pre>` element? Obviously, it's your site and you should make it how you want it to be, but I'll add that this makes it fundamentally pretty inaccessible. So you _might_ be aiding CURL, but you're cutting off anyone using non-visual browsing methods 😅
[tantek] Eg I can type example dot com like: http://example.com and Slack turns that into a hyperlink. I have no idea (don't care) of the plumbing it used to do so
btrem [tantek]: Not so sure I like that idea. That's taking a page advertised as text/plain and deciding it's something else. Slack is not, AFAIK, capable of html, so adding autolinking might make sense there. (Don't use Slack, so I don't know.) But a web browser is. If someone wants autolinking, then use what already works, text/html.
