I discovered Gemini in 2023 and really enjoy it. I like the minimalism. I like that page presentation is up to the user. I’ve even come to appreciate the line-based nature of its markup, because theoretically at least, it means a screen reader user knows that elements are always arranged top to bottom.

I say “theoretically” since I have yet to run into an accessible Gemini browser for Windows, which is unfortunate since Gemini’s whole deal is text first.

And to head off the inevitable comment about Linux, in order to have an accessible app you need an accessible OS for it to sit on top of.

  • tal@lemmy.todayEnglish
    3·
    2 days ago

    And to head off the inevitable comment about Linux, in order to have an accessible app you need an accessible OS for it to sit on top of.

    1. Boots up Android phone

    2. Installs F-Droid open-source package repository via apk

    3. Uses F-Droid to install Termux Linux environment

    4. In Termux, runs $ pkg install emacs

    5. In Termux, runs $ emacs

    6. In emacs, runs ESC x list-packages RET

    7. Scrolls down to elpher, gopher/gemini emacs client. Hits i for “install” and x for “execute”

    8. Runs ESC x elpher RET

    9. Starts tapping on links and browsing.

    Emacs runs on pretty much everything, so there’s a gemini client pretty much everywhere!

    • tal@lemmy.todayEnglish
      1·
      2 days ago

      I’ve never tried using emacs with my eyes shut, but aside from screenreader packages, I do see this:

      https://en.wikipedia.org/wiki/Emacspeak

      Emacspeak is a free computer application, a speech interface, and an audio desktop (as opposed to a screen reader). It employs Emacs (which is written in C), Emacs Lisp, and Tcl. Developed principally by T. V. Raman (himself blind since childhood, and who has worked on voice software with Adobe Software and later IBM), it was first released in April 1995.[1] It is portable to all POSIX-compatible OSs. It is tightly integrated with Emacs, allowing it to render intelligible and useful content rather than parsing the graphics (hence it is sometimes referred to not as a separate program, but a subsystem of Emacs); its default voice synthesizer (as of 2002, IBM’s ViaVoice Text-to-Speech (TTS)) can be replaced with other software synthesizers when a server module is installed. Emacspeak is one of the most popular speech interfaces for Linux, bundled with most major distributions. In 2014, Raman wrote an article describing how the software’s design was impacted by shifts in computer technology and its general usage over 20 years.[2]

      Emacspeak achieves its integration by being written largely in Emacs Lisp using “advice”, enabling it to literally be a wrapper around most functions that change or otherwise modify the display. Auditorily, verbalizations are pre-emptible, and common actions like opening a menu or closing a file have a brief sound associated with that particular action; it also immediately verbalizes all insertions of characters, and attempts to speak as much of the context sentences around the cursor’s present location as possible.

      Emacspeak facilitates access to a wide variety of content, from the web to DAISY books.[3]

      It does say “Posix-compatible”. Someome apparently did an emacspeak Windows port, but it hasn’t had updates for ten years, and it doesn’t sound incredibly mature:

      https://github.com/SaqibS/emacspeak-on-windows

      Emacspeak on Windows

      Emacspeak is an auditory interface to Emacs. It intercepts all Emacs commands, and produces feedback using text-to-speech (TTS) and audio cues (earcons).

      While Emacspeak is itself cross-platform, being written in Emacs Lisp, it uses speech servers to communicate with hardware/software speech synthesizers. To get it working on Windows, I have written a speech server for Windows, using the platform’s native TTS/audio APIs. The server is actually just a command line program which reads commands from standard input, as defined in this spec.

      This first version of the TTS server was really a proof of concept so that I could use Emacspeak more. Depending how much I use it, and how much interest I get from others, I may spend the time to make it more robust. Of course, pull requests are welcome. In particular, using multiple voices is not currently implemented.