Re: THEAD & TFOOT for columns

Jeni Tennison (jft@Psychology.Nottingham.AC.UK)
Fri, 15 Aug 1997 10:44:38 +0100


Message-Id: <l0302090db019cbb675c5@[128.243.31.122]>
In-Reply-To: <v03102827b018fc6fa3dc@[205.149.180.135]>
Date: Fri, 15 Aug 1997 10:44:38 +0100
To: www-html@w3.org
From: Jeni Tennison <jft@Psychology.Nottingham.AC.UK>
Subject: Re: THEAD & TFOOT for columns

At 19:27 +0100 14/8/97, Walter Ian Kaye wrote:
>Having to scroll in two directions is extra work for the user;
>limiting scrolling to only one direction cuts the work in half.
>And vertical scrolling is much more "natural" than horizontal
>scrolling. I think this is based on gravity; if you go back in
>time to when scribes published on actual *scrolls*, the scroll
>orientation was usually vertical -- you could drop the lower
>scroll to the ground and then read whilst operating only the
>upper scroll. Try that with horizontal papyrus and your neck
>will hurt. ;)

I agree that vertical scrolling seems more natural than horizontal
scrolling.  And in most cases, it is sufficient because of the fact that
text can wrap to fill the space available.

However, within tables with a lot of columns, or necessarily wide columns
(that hold an image, for instance), you're going to need to scroll
horizontally, or suffer with very thin columns of text that aren't
necessarily easy to read.

At 20:09 +0100 14/8/97, Neil St.Laurent wrote to the www-style list:
>Horizontal scrolling is great for:
>-accounting ledgers
>-time sheets
>-other charts
>-columnar display of multiple databse records
>-natural looking tree displays

... and in most of these it would be nice to have either the rightmost or
leftmost column being static.

Horizontal scrolling is already a reality for those of us that don't use
massive monitors, and it'll probably become a bigger issue with people
increasingly using electronic organisers, which have even smaller screens,
to 'surf the web'.  It is a particular issue with tables, because the
minimum width of tables is a lot larger than it is for most other things.
That's one reason why COLHEAD and COLFOOT might be a good idea.

The second reason is structural:  tables don't only have headings at the
top of columns, they also have them at the beginning of rows.  If a
facility is available to denote which rows make up a header and a footer,
there should also be a facility to denote which columns contain heading
information.  Any table could, almost by definition, be rotated such that
the columns become rows and vice versa.  To me, it therefore makes sense
that anything we can say about rows we should also be able to say about
columns.

Jeni

Jeni Tennison
Department of Psychology, University of Nottingham
University Park, Nottingham NG7 2RD, UK
tel: +44 (0) 115 951 5151 x8352
fax: +44 (0) 115 951 5361
url: http://www.psychology.nottingham.ac.uk/staff/Jenifer.Tennison/