Slide 1:
 
Slide 2:

            The hist() function draws a histogram
           
Slide 3:

            The VI() function from the 'BrailleR' package 
            can generate a text description of the histogram.

            In combination with a screen reader, this provides some
            information about the histogram for blind or visually
            impaired R users.
           
Slide 4:

            The hist() function returns an object containing information
            about the histogram as well as drawing the histogram.

            VI() uses that information to create a text description
            of the histogram.
           
Slide 5:

            The plot() functions does not return ANY information 
            about the plot that it draws.
           
Slide 6:

            This makes it hard for VI() to do anything.
           
Slide 7:

            The 'BrailleR' package has its own versions of some 
            major plotting functions that augment the information
            that is returned ...
           
Slide 8:

            ... and the 'BrailleR' package has some new 
            plotting functions ...
           
Slide 9:

            ... that generate information about the plot.
           
Slide 10:

            The 'BrailleR' package has to do a lot of work 
            for each individual plot that it wants to support
           
Slide 11:

            'ggplot2' is a quite popular package for drawing plots in R
           
Slide 12:

            But 'BrailleR' could not generate text descriptions for 'ggplot2'
            plots
           
Slide 13:

            'ggplot2' always generates a "ggplot" object
           
Slide 14:

            And 'ggplot2' exports the ggplot_build() function, which
            generates more information about what is drawn
           
Slide 15:

            So now 'BrailleR' can not generate text descriptions for 'ggplot2'
            plots!
           
Slide 16:

            And this just works for a range of plots straight away
           
Slide 17:

            And this just works for a range of plots straight away
           
Slide 18:

            Deciding what to say and how to say it is hard;
            there is no "best" solution
           
Slide 19:

            So we allow for flexibility;  people can use their own
            template (based on the R package 'whisker' and the 
            Mustache templating language; ...
           
Slide 20:

            ... and people can do things with the object, like searching
           
Slide 21:

            ... and people can do things with the object, like sorting
           
Slide 22:

            The overall design involves gathering information about
            a plot, generating text from the information (using a
            template), and then writing methods that work with the
            combination of information and text.
           
Slide 23:

            With 'ggplot2', we describe a plot in terms of a small
            number of building blocks - we can create an infinite
            array of plots by combining the building blocks in
            different ways.
           
Slide 24:

            This means that the support for 'ggplot2' in 'BrailleR'
            can just add support for a small set of building blocks
            and then we can generate accessible text for an infinite
            array of plots.
           
Slide 25:

            Slide says it all
           
Slide 26: