From bbdff25094ee114ec2dad76195896d0c181c7036 Mon Sep 17 00:00:00 2001 From: Adrian Cochrane Date: Wed, 8 Apr 2020 13:47:53 +1200 Subject: [PATCH] Add new ISSUE --- ISSUES/in-progress/links.md | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) create mode 100644 ISSUES/in-progress/links.md diff --git a/ISSUES/in-progress/links.md b/ISSUES/in-progress/links.md new file mode 100644 index 0000000..53eadcc --- /dev/null +++ b/ISSUES/in-progress/links.md @@ -0,0 +1,25 @@ +# Links + +You should be able to follow links by repeating back it's text either via speech or typing. + +--- + +To expand upon this design, if only a single link matches the browser would immediately start loading that page and announce that it's doing so. +But if multiple links match it'll list each link's text, title, and href alongside a shorthand letter with which to refer to it in your response. + +--- + +These links would be defined very generally, and include: + +* "psuedolinks" inserted by the browser to take the place of toolbar buttons. +* "petnames" used to quickly navigate to prespecified sites. +* bookmark "tags" used to search your bookmarks. Or should I just rely on petnames here? + +These different types of links would be styled differently during the clarification stage so users are clear what's a part of the browser and what's part of the web. This distinction may be necessary for security, and I'll struggle to maintain it otherwise. + +--- + +- [x] Extract links from page +- [x] Switch to C for access to needed libraries +- [ ] Filter links based on user input +- [ ] Integrate voice recognition (CMU Pocket Sphinx or Mozilla Deep Speech) \ No newline at end of file -- 2.30.2