[bramd]

> Though sadly it will be a while until we will hear: "You are visiting Hacker News, in a simple look and without any images, it features a list of links to articles and a menu at the top. Do you want me to read the articles titles, continue with accessing the menu or a more detailed description of the page?"

Sadly? I strongly disagree. If all my human-computer interaction would be like communicating with Siri/Google Now (e.g. conversational), I would have ditched computers years ago. It's all about giving semantic information to screenreaders so the user can eficiently navigate that content. In the rare case that I need some kind of human description, I just share my screen with someone that has working eyeballs.

FYI, I'm totally blind.

[anotheryou]

oh, great to hear from someone actually needing this. But than I'd still argue, that semantics should be possible to parse from the style of the site. Big letters? Must be a headline.

Well actually I would love html to be more to be like latex/markdown, I just thought, that if I recognize it as a list, it should be parsed by a screen reader as such too.

May I ask you which screen reader you use and at what speeds? (or do you use some braille screen?)