Vivien (dentelle & numérique)Right now I like the #indieweb #webmentions / #microformats more than #activitypub / #jsonld, because the author chooses which replies appear below their post. I think it would be better if there was an expectation to regularly validate the webmentions (we have HTTP caching for that), and also for #xhtml instead of html. I’m not familiar with the LD markup languages, but maybe there exists something more interoperable than microformats (with proper URIs for properties).
Khleedril@khleedril As a webmention implementer, I need to parse the markup language. However, contrary to XHTML, it is near impossible to parse HTML correctly (a certain anti-anti-evil company is making sure it stays that way, to protect its monopoly on web clients). Our best approximation is the almost-HTML parser in libxml2, but it risks going unmaintained.
I don’t think I need to expand on how building on standards is a requirement to avoid massive security issues.
I don’t think I need to expand on how building on standards is a requirement to avoid massive security issues.
@gugurumbe I personally use the #Rust scraper crate to do this: you basically tell it to extract sections by CSS path, and it always does the right thing. I think sometimes lax standards are better than rigid ones, and that current HTML/CSS serves the #indieweb very well. X{HTML,ML} is all too finicky for the casual web producers like me.
[ps. there is also a Rust microformats crate, but that's been rubbish in my experience.]
@khleedril yes, depending on a web browser solves the parsing problem, but unless it’s Google’s, it won’t be correct. I would rather write my own parser for a HTML-inspired language.
The difficult part is when two different, slightly incorrect implementations understand the document differently. This is how we get security flaws.
I understand that most HTML authors do not care much about their documents, but when machines are involved (such as microformats parsers), we need precision.
The difficult part is when two different, slightly incorrect implementations understand the document differently. This is how we get security flaws.
I understand that most HTML authors do not care much about their documents, but when machines are involved (such as microformats parsers), we need precision.
Ogey Rrat@gugurumbe @khleedril None of the browser vendors want to use XHTML that's why WHATWG was made when W3C wanted to replace HTML with XHTML.
@Unn0wn Noone would be surprised, should I declare I don’t have much respect for the browser vendors [1 is less known than other complaints]. But that’s beside the point. technology is political, so we have to think about our choices at every step. Regarding the specific task of parsing microformats, using HTML gives more power to a monopoly, while XHTML does not.