Some questions about data tables

I have some questions concerning the design of data tables. I was wondering
if anyone would be able to give me any answers on them? There are three -

1 What is the proportion of screen readers that are currently capable of
associating headings and data cells by reading the mark-up of a table?

2 - single and two level tables

single level tables  - table follows

Bus Times for different days:

Saturday      9    12   13
Sunday       10   13    14

According to the accessibility guidelines all that is required to make this
accessible is to mark up the headings properly, it is not necessary to use
'scope' attributes or other tabular mark-up (Guideline 5.1 & 5.2). The
accessibility guidelines do not mention the need to ensure the linearisation
of data tables, only layout tables (guideline 5.3). Am I right in assuming
that this table will satisfy the accessibility guidelines with the correct
headings marked up?

2 level tables - table follows

Bus Times, columns refer to places, rows refer to days of the week

Day           London  Manchester    Leeds
Saturday       9             12                  13
Sunday         10           13                  14

The accessibility guideline 5.2 states that tables with 2 or more logical
levels should be marked up to indicate the logical relationships between
cells. So here I assume that by marking up both column and row headers with
'scope' attributes, the necessary relationships would be in place.

The problem I have here is that I don't know the level of support of current
screen readers for the 'scope' attribute, or the other attributes added by
the HTML 4 specification. Does anyone know of any research that has been
done in this area?

3 - Automatic assumption of relationship between headings and cells

Are there any screen readers which are capable of associating headings by
assuming an association with the column or row to which it refers without
any extra mark-up? It was my thought that in most cases the association of
headings and cells was self-evident and so it could be automatically
assumed. The heading in the corner of my table would cause some confusion as
the relationship to cells is not self-evident there. However assuming the
association in the other cases would give some extra information
automatically. Any thoughts?Accessible generation of tables -

Received on Wednesday, 22 December 1999 18:49:48 UTC