Neils, I think what you are about to find is that individual preference generally dominates accessibility concerns here.
>>>>> "Christian" == Christian Schoepplein <[email protected]> writes: Christian> Hi Nils, Christian> On Sun, Jan 28, 2024 at 11:54:24AM +0100, Niels Thykier wrote: >> In my case, my tool is not interactive. It generates text output >> and either emits all of it on standard out or pipes it to a >> pager. It is closer to a tool like ls than mutt in spirit. As far >> as I know, my tool cannot control the caret of the output in any >> meaningful way in this case. Christian> Piping output into a pager is very uncomfortable for Christian> screen reader users IMHO. To navigate such content you Christian> either have to use the keys of a connected braille device Christian> or use the screen reader commands to go through the Christian> output. Both variants are much less good then piping the Christian> content into a programm where the user can navigate the Christian> output just with the normal arrow keys, like navigating Christian> through a file in an editor, or piping the content into a Christian> textbased browser like w3m where the content also can be Christian> navigated by using the arrow keys. I'd much rather have content in less than in w3m. 1) I do have my screen reader controls optimized for something I like to use, so being able to navigate with the screen reader keys is in my mind a plus not a minus. 2) I always have less available; I would not normally expect a text mode browser to be available on a remote server. 3) I find pagers to be a good interface in the common case I actually want to read the entire screen. I find the simple ":" prompt of less or the "--more--" of more less disruptive than a status line of an editor or browser. For me, reading tabular output, the best advice I can give is to focus on making the table row elements as simple as possible. For example separating each row element with a single pipe and spaces rather than something more complicated. I appreciate that the header line--or the separator between the header and the rest of the table--may be more complicated. That's not a big deal, but the more lines with lots of dashes I have to work through, the more annoying. I would definitely find pipe separated text easier to read than CSV. Although in exceptionally complicated tables, I would find being able to take CSV into a spreadsheet convenient. For reasonably simple tables, I would probably choose to read pipe separated text rather than to take HTML5 and read it in Firefox. I find the text output of sqlite entirely reasonable unless the number of columns in the table gets large enough and the data is sparse enough that I cannot remember what column is involved. I generally refine my query in that case rather than turning the table into html. --Sam
