This is a reasonable question, Kenny. Let me start by agreeing that, all things being equal, there's nothing to be gained. But, all things aren't equal. For example, it takes much more in the way of system resources to run gnome with reasonable performance. The difference isn't slight, but quite large. For the blind user, those are all wasted resources, wasted on painting a gui that we don't see. Another example is portability. I'm expecting several devices will become available in the next year or two that run Linux--portable devices that you carry around like you might carry a Braille Note today. These will be ncurses capable, but will undoubtedly not be gnopernicus capable. Lastly, there's the inequality of the AT itself. Of course someone should continue working on gnopernicus, and someone will. I'm not sure there has been much "we" in that process, despite the open license, for various reasons. I'm not sure how much "we" there will be going forward either, for various reasons. But, I am sure there will be continued development until a robust, reliable, and effective screen reader for the gui emerges. I'm sure there will be such a thing, and I'm sure it will read QT4 based apps as well as GTK+ ones, because the fat cats need such a screen reader to keep selling to our Uncle Sam. Now, I will confess to a certain degree of impatience. However, that's not the whole story. Most to the point, I don't believe it would be necessary to contrive a ncurses interface for each individual app one wanted to port. If that were so, I would judge such a suggestion futile. The beauty of at-poke resides in the fact that it is generic. It simply generates a report. I believe that "report" could be turned into a generic, interactive menu. Remember that there is a gnome AT tool called gok, the gnome on screen keyboard. It's designed to accept input, where gnopernicus is designed to deliver alternative output. Gok can take input from all kinds of mechanisms, simple serial switch on up. And, it handles them very intelligently. Unlike gnopernicus, it is a gnome accessibility success already. Kenny Hitt writes: > Hi. Can you provide examples where writing a ncurses program to act as > the interface to a Gnome app would be helpful? Based on my experience > with Gnome, if the program provides the accessibility info, then it is > accessible in Gnome. If it doesn't provide the accessibility info, I > can't see how adding the extra layer of an ncurses app talking to a GTK > app will help. > I agree Gnome screen readers still need work, but wouldn't it be better > to focus on improving them? > > Kenny > > On Tue, Jan 11, 2005 at 09:09:59PM -0500, Janina Sajka wrote: > > Laura Eaves writes: > > > Hi Janina -- > > > I tried sending this once before but it bounced due to message size... > > > But as for gnome: > > > That is interesting about the query capabilities of the gnome interface -- > > > but I disagree that implementing a full ncurses, text-only representation > > > based on these queries is necessary, as that would just require a lot of > > > useless code to do screen layout of text when graphical screen layout has > > > already been done by gnome, right? Actually, some users don't know this, but > > dGo > > > > I didn't say it was necessary. I said it was possible and that there are > > good reasons to consider doing it that way, not the least of which being > > that gnome is not very accessible, whereas the console is. > > > > _______________________________________________ > Speakup mailing list > Speakup at braille.uwo.ca > http://speech.braille.uwo.ca/mailman/listinfo/speakup -- Janina Sajka Phone: +1.202.494.7040 Partner, Capital Accessibility LLC http://www.CapitalAccessibility.Com Chair, Accessibility Workgroup Free Standards Group (FSG) janina at freestandards.org http://a11y.org If Linux can't solve your computing problem, you need a different problem.
