On Sun, 3 Jan 2016 06:22:11 +1300, Chris Bannister <cbannis...@slingshot.co.nz> wrote:
>> You would do well to read *all* of Steve Matzura's posts before >> bemoaning your lot. You'll come across "speech synthesis" and >> "screen reader". > >In that case, mails in html must be almost impossible to comprehend. :) >I'm guessing there is some sort of configuration available, so something >like t-prot (apt-cache show t-prot) could be used. I'm only guessing >here, I've never had any experience with a screen reader. Just for grins, here's the two-minute lesson. By and large, a screenreader speakes what it sees. Applications like Microsoft Windows Live Mail and Outlook (the Express and real versions), Mozilla Thunderbird, and probably one or two others, render electronicmail in HTML format quite well, just like browsers render Web pages, denoting the presence of links and other such controls definable in HTML. Otherwise, as long as it's text, real text, not a picture of text (like a scanned document or picture of, say, a sign or a book cover), screenreaders handle it nicely. There are even screenreaders now that have built-in OCR for such exception cases as just mentioned. They know they should speak and track things that are in a different color than the rest of the text on a screen, which means they can track highlighted portions of text as a cursor moves or a selection bar changes in combo and list boxes, they report the status of checkboxes and radio buttons, multi-select list boxes, buttons, all the standard Windows controls, of which I think there are thirty-five. Where things go off the rails for screenreader users is when application developers use non-standard Windows controls or navigation schemes, or disable TAB-navigation entirely, because then the screenreader has no reference point of what it should be speaking. This is particularly annoying, not to mention frustrating, when changing a control on a screen causes the whole screen to update, and once again, the screenreader loses context, so, calling on its fine command of language, it says nothing, and the user never knows what happened. :-) Oh yes, screenreaders know emoticons and some emoji, too.
