On Wed, 21 Feb 1996, Tom Magliery wrote:
> We're implementing the new TABLEs model in NCSA Mosaic. The TBODY will be
> a scrollable region in the table, with THEAD and TFOOT fixed. It will
> probably look something like this:
[... nice drawing deleted]
> My question is, what are people's feelings about deciding how tall the
> TBODY should be?
> Let dwh = document window height, thh = thead height, tbh = tbody height,
> tfh = tfoot height; and ignore details like widths of borders, margins and
> whatnot. Here are some possibilities for tbh:
> 1. dwh - (thh + tfh)
I also think this is the best basic choice. See more comments below
> 2. height of N tallest row(s) in TBODY
> 3. some function of thh,tfh independent of tbh. For example, maybe tbh =
> Each of these has problems:
> (1) fails if THEAD and TFOOT are too tall, but seems pretty good otherwise.
> There's also the question of what criteria establish that THEAD and TFOOT
> are "too tall". (Probably a percentage of dwh.)
I would define too tall as the tallest row won't fit in tbh as defined
above. Given choice (1), I'll try and address the 'edge' conditions:
a) The whole table fits in dwh -- obvious ... show it all
b) tbh won't hold the tallest row -- my choice would be for graceful
degrading by first sliding the tfoot down until just enough remains
in the window (e.g., 1/2 of a line) to hint that the foot is there.
Then slide the thead up until just a hint remains at the top (keep
last line of the thead?). If the trh (tallest row height) still
doesn't fit, just scroll the body.
THe % of head/foot to anchor in the window seems a good candidate to
control via style sheet and user preferences.
You could also fall back to rendering the table to fit the height of
dwh if the UA window were maximized.
Having decided how high to render the table, document scrolling should
make it easy for the user to see the partially suppressed thead/tfoot.
Also, it would be nice if the user could resize the table
c) If dwh is too small to display the hints in (b) + scroll bar for tbody,
then your guess is as good as any.