Handheld device for partially-sighted and blind people
Because blind people don't need a screen, and even many partially-sighted people use text-to-speech rather than a screen for some tasks, it's possible to provide for these people a computer that is closer to a tablet than a laptop in portability, while still offering a comfortable keyboard for rapid and prolonged typing. Especially when combined with the extended battery life that comes from having no screen, such a computer would be especially appealing to blind and partially-sighted students for use in the classroom.
For maximum portability, this housing should be based on a braille keyboard. To be clear, we don't mean a keyboard with braille on the keycaps. Rather, we're referring to a keyboard with keys corresponding to the dots in a braille cell, plus a space bar. For 8-dot computer braille, that makes a total of 9 keys. With so few keys, a braille keyboard can be quite compact while still being perfectly comfortable. Of course, the reason why so few keys are necessary is that one presses multiple keys at once to input each character. To go beyond text input and enter commands, the standard practice is to combine one or more of the dot keys with the Space bar; these keystrokes are called "chords", and there are many de-facto standard chord commands.
For an example of how compact a braille keyboard can be, see the BraillePen Slim. (Note: Ignore the section about the BraillePen 12 Touch, which includes a refreshable braille display that is well beyond the scope of this project.) See the picture of the BraillePen Slim to get a general idea of how a braille keyboard is laid out. The BraillePen Slim measures 148 x 67 x 21 mm, or about 5.8" x 2.6" x 0.8". Obviously, a full-featured EOMA68 computer housing with a reasonable battery, ports, and speakers would need to be larger; the dimensions of the BraillePen Slim simply show us a lower bound on how compact a braille keyboard can be.
The main challenge in producing a computer housing based on a braille keyboard would be the development of microcontroller firmware and/or user-space software for translating braille keystrokes into something that a general-purpose operating system and applications can handle. It should be feasible to implement microcontroller firmware to translate braille keystrokes to QWERTY equivalents if we limit ourselves to computer braille input. However, proficient braille users expect to be able to input contracted braille, such as grade 2 braille in English, in which one braille keystroke does not correspond directly to one QWERTy keystroke, and in fact, braille input needs to be translated into normal text a word at a time. For this, we'll almost certainly need a user-space software component that integrates tightly with the chosen desktop environment (probably MATE), and perhaps also with some applications such as LibreOffice Writer. Fortunately, the tricky business of translating from contracted braille into normal text is already handled by an excellent free software library called liblouis.
As for the rest of the hardware in the housing, since audio will be the sole means of output, a USB audio codec, at least a mono speaker, and a headphone jack are indispensable. There should also be at least one USB port and a microSD slot, as in the laptop housing. An internal compartment for a USB WiFi dongle, as in the laptop housing, would be nice, though that would take up more space. Ideally, the battery would have enough capacity to provide 8 to 12 hours of battery life without a screen.
Who is behind this idea?
This idea was proposed to the arm-netbook list by Matt Campbell, who has several years of experience as a software developer in the field of assistive technology for blind and partially-sighted people. Among other projects, Matt has developed a screen reader for Windows, and has developed software for a proprietary PDA for blind people.
Discussion on arm-netbook: http://lists.phcomp.co.uk/pipermail/arm-netbook/2016-September/012053.html