- From: Will Pearson <will-pearson@tiscali.co.uk>
- Date: Thu, 11 Nov 2004 17:56:31 -0000
- To: <www-svg@w3.org>
Jonathan Chetwynd wrote: "> Will, > > I'm completely with you in respect of the need for good examples, and > would welcome contributions. > Do you have any? would you be willing to create some for discussion? > > I'd welcome techniques for creating accessible SVG1.1 content, so don't > think that is a valid reason for delay*. > Any techniques document will take a considerable time to author, > comment on, reach a consensus, draft and publish. > almost certainly years rather than months. There's always a ready > excuse for delay." Jonathan, I'm not delaying in respect to groups for which SVG is truely accessible currently. I would be quite happy to help in any endevours to promote SVG accessibility for these groups. However, I don't believe the world of diagrams and images are truely accessible for all at present. It could be, particularly if directional navigation was based on visual spatial relationships, not element order, as this would allow non visual users at least some access to the semantic meaning encoded by spatial relationships within diagrams, drawings, images, etc. > Please don't delay instead of verbally describing an issue, which you > have done well, why not create, or point to a specific example? > I have no doubt that this will greatly enhance everyone's understanding > of this issue. Being blind, and thus unable to draw at present, I'll create a description, as that's a form accessible to me. Please excuse the referencing, as it's come from a research paper of mine on semantic graph transformation using SVG. Line graphs are a common form of complex data type, and one that blind people will come across in education and daily life. At present there exists designs for presenting the data contained within a line graph to the user (Mansur D. L., Blattner M., and Joy K., 1985; Edwards A. D. N. and Stevens R. D., 1993; Brown L. and Brewster S. A., 2003), however this body of work does not specify how the numerical information contained within a line graph can be extracted from a visual diagrammatical form. SVG, as explained earlier, allows for identification of shapes and their locations, and this process can be used in extracting the numerical values that are to be presented to a blind person. It will be assumed that the textual labels conveying the values for each of the two axes will be drawn using the SVG text tags, and that the line conveying the numerical values for the plot will be drawn using the line tag. The labels for the axes can be obtained by examining the text tag, with the spatial position identifiable through examining the associated X and Y attributes. For example, if the x axis contained three labels for 1000, 2000 and 3000, they could be represented in SVG through the use of the following tags: <TEXT X="5" Y="5">1000</TEXT> <TEXT X="10" Y="5">2000</TEXT> <TEXT X="15" Y="15">3000</TEXT>. The fact that these labels form the x axis can be identified by the fact that they all have the same value for the Y attribute, if they were Y axis labels, then the X attribute value would remain constant whilst the Y value would vary between tags. The association between the value of the label e.g. 1000, 2000 or 3000, and the coordinate is important. This association between value and coordinate specifies that any line at the X coordinate of 5 will have the value of 1000, the X coordinate of 10 the value of 2000 and so on. This association provides a clear method of identifying the value of the line at given points on the axis by comparing the line coordinates to those of the text tags for the labels. If a line graph were drawn using the following tags: <TEXT X="5" Y="5">1000</TEXT> <TEXT X="10 Y="5">2000</TEXT> <TEXT X="15" Y="5">3000</TEXT> <TEXT X="1" Y="5">10</TEXT> <TEXT X="1" Y="10">20</TEXT> <TEXT X="1" Y="15">30</TEXT> <LINE X1="1" Y1="1" X2="5" Y2="5"> It could be identified that the line originated at the origin for the graph through the value of 1 for the X1 and Y1 attributes, and that it finished at the value of 1000 on the graph's x axis and 10 on it's y axis, by examining the X2 and Y2 attribute values, and comparing these to the coordinates for the <TEXT> tags that form the axis labels. --- end of example --- I can't see a way for alt to resolve the complex nature of the spatial relationships between the lines forming the value, and the lines forming the axis. This is based on spatial relationships between specific points on several <LINE> tags, and thus one single alt attribute per element won't suffice, it would need to be one alt attribute per point on that line. Will
Received on Thursday, 11 November 2004 19:34:44 UTC