💾 Archived View for idiomdrottning.org › re-a-reader-friendly-browser captured on 2024-08-25 at 00:58:58. Gemini links have been rewritten to link to archived content

-=-=-=-=-=-=-

Re: A reader friendly browser

Gerikson writes:

We had that dream, in the early 2000s. It was called The Semantic Web

I think that’s one reason why Gemini is succeeding. (It’s already as much of a success as it needs to be. Smol. Anything else that happens is gravy.) With fewer line types and a narrower grammar for semantics, it’s less easy to mess it up and there’s less data to scrape and exploit.

I’ve ranted before on how many Markdown sites will have you render <em> for italic things that aren’t emphasis at all, like titles and foreign names.

The semantic web failed more because people were not able to correctly express themselves using all the nuance that the semantic web’s specs demanded. It was like forcing people to learn a whole new meta-language that included invisible nuances of typography (“whaddayamean <cite> different from <em>? Look the same to me?”).

Gemini has comparatively fewer things to learn (seven line types, three of which are optional) and comparatively fewer ways to mess up:

header lines in the wrong order
header lines for emphasis instead of for denoting structure
lack of alt text on pre blocks
ascii-clobbering of texts (like asterisks around words or bracketed numbers throughout the text)
hard wrap
missing or wrong lang headers

Maybe a few others, compared to the one million ways to mess up the semantic web.

(Also some client bugs like how some versions of Elpher has a really hard time with emoji in links.)

Mostly, people can do it, and the consequences of messing up on these are pretty mild, while the semantic web’s entire reason for being depends on people dancing exactly in line.

Matthew writes:

Browser support for Markdown, epub, and a variety of other user-friendly formats

I love Markdown, but it’s not a good exchange format, since people’s preferred spec and implementation vary so wildly. It’s a great authoring tool locally that can be used to translate text into a more robust exchange format.

I love Markdown

Better client side support for overriding CSS is a good goal, and already we have things like Calibre, which does a great job of mangling web pages into ebooks, and the Reader mode in browsers like Safari, and browsers that don’t even look at the CSS, like Lynx or edbrowse.