Hello everyone,
My name is Rafael. My college Danilo and I are working to improve evince accessibility as our Monography for our conclusion in the Computer Science Course At Senac University in Brasil.
We have one year to develop this job, the current semester and the next one. We have already spoken with some developers, particulary Nikolay who was very atencious with us, giving us importat ideas.
The fact was that, altough we had previosly worked with the C programming language, we had no experience with Gnome, autotools, GTK, GAIL and ATK. The last two we hadn't even heard about when we decided our subject.
We realised that we couldn't produce an artifact in this semester, so we decided to make the most of the oportunity to study those architetures.
We learned to make an interface in GTK, for this we used the Glade software. We had to use the old version because the new one doesn't export code to C, now it works creating an XML file that the running program loads.
There we saw how GAIL works, how an screen reader software as ORCA can get information from the GTK labels, and in the case they don't exist it takes the data from the ATK Label. We could made an implementation of a skeleton of an Gtk program with ATK interaces implemented with a dummy code.
But this kind of accessibility improvement evince alread has. I think that we will concentrate in two new features, such as the full keyboard navigation and an ATK interface that would let the pdf text be read by orca, for example.
What we produced this semester was some documentation in portuguese (we are from Brasil), explaining with examples (kind of step by step) what we learn about Gnome, Accessibility, Version Control, Autotools, Glade, Orca and Accerise (accessibility browser software that helped our understanding). We also described how to get Eclipse working with CDT and GDB and some further steps to work with version control). This can help future portuguese speaking developers to spare some time.
The July month is winter vacation here and we will have some more time to work on the code (both of us work and can't expend as much time as we wish on the project).
So, we would like to know if someone can explain a little bit more (or if there is already a document) , how does the evince architeture works, because just debugging the code we couldn't understand everything.
We know that evinces works with the poppler. Nikolay told us some ideas he had. Implement a kind of html interface between evince and poppler, that could allow more complicated things, such as text with coordinates. What we got from the idea, is that while poppler is rendering the text, it should also feed n ATK text interface in evince that would work in parallel with the texto box to implement those functionalities.
Resuming, we don't have an specific doubt at this time, but any opinion would be very aprecciated because we are still understanding our next steps. We have joined the Evince Mailing list so we can catch up with the current discussion.
Thanks to everyone....
Rafael and Danilo