unshorn-forks / BrailleR

An add-on package for R primarily aimed at improving access to R for blind users. The package contains tools and functions that will help get a student through their first year statistics course, and probably well beyond it.

The BrailleR package is a collection of tools to make use of R a happier experience for blind people.

The BrailleR package does have dependencies, all of which are available on CRAN. Version numbers are of the form a.b.c, where a.b is the major version and c is the minor version. The leading zero infers that this package is under ongoing development to get to the state where it delivers the original aims of the package. The second part of the major version is odd for development and even for CRAN releases. Minor version increments just help track progress.

Blind people's use of R

Blind people cannot use RStudio and depending on the combination of operating system and screen reading software we choose to use, we might have differing levels of success with the standard R GUI.

R is perhaps the most blind-friendly statistical software option because all scripts can be written in plain text, using the text editor a user prefers, and all output can be saved in a wide range of file formats. The advent of R markdown and other reproducible research techniques can offer the blind user a degree of efficiency that is not offered in many other statistical software options.

Specific objectives of BrailleR

1. Make accessing output simpler.
2. Gain access to the content depicted in a graph.
3. Make it easier for blind users to create their own R markdown documents.
4. Replace the tasks other users can do quickly using the RStudio IDE.

Installation of BrailleR

Use the following commands if using the CRAN version of BrailleR:

chooseCRANmirror(ind=1)
install.packages("BrailleR")

Or, if the CRAN version is not recent enough or your liking:

chooseCRANmirror(ind=1)
install.packages("remotes")
remotes::install_github("ajrgodfrey/BrailleR", upgrade=TRUE, quiet=TRUE)

You can help

Any assistance to fill in the numerous gaps in what BrailleR delivers will be most gratefully received. As the primary developer of this package is himself blind, attempts to find out if the tools being created actually do replace visual elements or tasks requires a sighted person to sit alongside to offer commentary.

Code of Conduct

Please note that this project is released with a Contributor Code of Conduct. By participating in the BrailleR Project, you are agreeing to this code of conduct. Please check you agree with its terms before participating.

Wish list

Some issues to resolve:

1. Add to the work done by Debra Warren and Paul Murrell on VI.ggplot() and related functions. Of critical importance is to expand the rane of geom_...() objects are described.
2. storing the graphical parameters in ScatterPlot() etc. does not necessarily get the right ones selected. Need to replace this by somehow keeping the ... arguments supplied by the user. Tested using col and pch in FittedLinePlot() N.B. possibly redundant if moving entirely to ggplot graphics
3. Get extra bits of information for WTF() coming through, such as other shapes, lines etc.
4. ascertain which shiny app widgets are useful for blind users' screen reading software.
5. link various plot objects to sonify package.
6. get more graphics ready for making into SVG for use on Tiger products; mostly in SVGThis() method
7. Get more Graph types working with the additional XML to make them interactive.
8. get graph files named properly by TwoFactors(); including boxplots etc vs each factor not both.
9. In OneFactor() and TwoFactors(): ensure that boxplots are created for each factor unless a level is short of reps.
10. WTF() doesn't pick up multiple outliers in boxplot() as more than one point
11. fix VI.lm to cater for the weighted lm case.
12. check the css is working for VI.lm()
13. work out how to change colours in R console for low vision users as commands and/or settings
14. create function to re-create vignettes from Rnw source into HTML
15. Multiple language support was started but is in serious need of some love.
16. Functions for the VI method to be completed include: VI.prcomp(), VI.factanal(), VI.glm(), VI.htest()?
17. added functionality to compile all Rmd files in the current directory. Request from JYS; initial implementation done in June 2017. Includes batch file for use in Windows explorer and DOS prompt. Needs testing.
18. VI.scatterplot() needs creating N.B. possibly redundant if moving entirely to ggplot graphics