BookmarkSubscribeRSS Feed
RandyHerbison
Calcite | Level 5

Hi,

 

I'm considering using hidden text to make the data in images of charts accessible to screen readers such as JAWS and NVDA,  The report requirements preclude using visible tables, within the file or linked, for alternative presentation of chart data.  

 

Would layering image data behind an image fail any WCAG 2.0 success criteria?  If not, is there any reason why formatting the hidden text as an accessible table wouldn't be preferable to unstructured data?

 

Thanks

1 REPLY 1
gdkraus
SAS Employee

I’ve included my opinion below, however, please consult your resident 508 expert or the 508 officer at your client’s agency for their input.

 

Before answering the initial question, the most important question to answer is what is the most effective way to communicate this information. Are you able to provide a description of the graph to the user that accurately conveys the information instead of presenting the raw tabular data? Understanding the trends and patterns that a graph represents can be very difficult for a user to do by simply reading the raw tabular data, especially for tables with more than a few rows of data. In cases like these, a description stored as the alt text for the image can be much more useful and is the preferred solution.

 

proc sgplot data=myData description="graph description goes here";

 

If a non-visual table is required, there are a couple of things to consider. This issue relates to WCAG 2 1.1.1 Non-Text Content. Here are some useful resources.

The key concepts here are the following:

 

1. Can the text alternative be "programmatically associated" with the graph? In other words, can the user find this alternative content? This can be met by referring to the visually hidden table in the alt text of the graph, using the DESCRIPTION option of SGPLOT. This way the user can know what and where the alternative content is.

 

2. Can the text alternative be "programmatically determined"? In other words, can the alternative content actually be used by the browser and assistive technology? I think this can be met in most cases using a visually hidden table, because screen reading software can access and navigate this table. There is one caveat to this - it may be a violation of Section 508 302.2, Functional Performance Criteria for Users with Limited Vision. This would be a case where a user might need to magnify the screen and they have a screen reader running to assist them (ZoomText does this), but when they encounter this graph they won't have any way to match up what is visually presented on the screen with what the screen reader is reading to them.

 

The technique you are describing is not a common implementation, where you would visually hide significant portions of content on the page. Given the limitations of what is possible technically in PDF, the solution you suggest may be the best alternative in some circumstances, where a table of data is the only way to accurately describe the graph and the graph alt text on its own cannot suffice.