Re: operational concept for table browsing

On Wed, 1 Oct 1997, Al Gilman wrote:

> I think I need to articulate some of the assumptions I am
> operating under.

Good idea.

> One is that, while it is well and good to talk about how one
> would do a full readout of a table under different sub-class
> hypotheses, that interactive browsing of the table is more
> critical to successful use by speech users.

Absolutely. I feel it is critical to enable browsing, including
the ability to suppress and expand levels of detail.
 
> I have been assuming that HTML needs to have enough semantic
> content to fully support the following operational requirements
> for table browsing.

> 1.  Cell focus.  The user needs to be able to focus on any
> individual cell.  That means that the content of a cell can be
> read out in entirety without interference from content from
> outside the cell.
> 
>  - This is adequately supported in the HTML.  It may require
>    some enhancements to communication between browsers and
>    access adaptors.

To meaningfully focus on a cell in isolation, I believe there
is a requirement to be able to state the headers that describe
this cell. This further leads to a need for abbreviations for
headers.

> 2.  Cell navigation.  The user needs to be able to move the cell
> focus around in the table.  The minimum capability is discrete
> stepping through the cells independently in row and column
> dimensions.

This is overly limiting. Why force a blind user to conceptualize
a visual model for the table, where you have to think in terms of
moving up/down or left/right? By providing a way to consider the
table in terms of a set of axes where related headers label
positions along these axes, we can get around a particular
visualization of this. This is particularly important when we need
to restructure the table, e.g. for Braille with an extremely small
numbers of characters per line.

>  - This is adequately supported in the HTML.  It may require
>    some enhancement to the command repertory of browsers.
> 
>  - Enumerators use for row/column coordinates may be
>    a topic for styling language.  Go-to-cell by coordinates
>    may be supported but is beyond the minimum requirement.

An alternative view point is to consider the table in terms
of a number of axes for grouping data independently of visual
realizations. This leads to a simple way to interact with tables
aurally in terms of walking a hierarchy. It allows you to suppress
levels of detail, which seems essential for large tables. 

A simple CSS property can be used to influence how each node is
spoken.  Do you say all of the headers above this point in the tree
or do you suppress them until the position along one of the axes
changes? This feature is included in the latest CSS2 draft. 

> 3.  Cell explanation.  The user needs to have the option to be
> presented a definition of the variable that the cell contents
> evaluates.  Compare with the colloquial "If that's the answer,
> what's the question?"  This is analogous to the normally-hidden
> cell formula in a spreadsheet.

One idea is the ability to describe the cell in terms of the headers
that describe it. However, this task should be seen in the wider
context of browsing the table as a whole. Another is the use of the
title attribute to provide a justification for the cell's data.

>  - This requires new provisions in HTML.

Axis/Axes are old ideas and date back several years. I have refined
them based upon further study and wish to now add a scope attribute
to simplify common cases. This provides a simple way to specify
values for which cells a header applies to. 

Scope is one of: "row", "column", "rowgroup", "colgroup" or
implicit. It may be provided with the following elements:
THEAD, TFOOT, TBODY, COL, COLGROUP, TH and TD. It indicates the
cells to which this header applies. This can be overridden by
specifying an explicit value for axes. Scope encourages authors
to structure the table into row and column groups. 

I would like to cover scope in the next draft as time is now
pressing.

Regards,

-- Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
phone: +44 122 578 2521 (office) +44 385 320 444 (gsm mobile)
World Wide Web Consortium (on assignment from HP Labs)

Received on Tuesday, 7 October 1997 10:02:01 UTC