Re: Changing the name for the table AXES attribute

to follow up on what Dave Raggett said:

> 
> If we want to change the name from axes, I strongly
> recommend using a name that immediately reflects the allowed
> attribute values. As these are generally headers it follows
> that the name should be "headers". In my opinion, keep-with
> is just as obscure as axes and would confuse authors.
> 
> I therefore propose we rename axes to "headers"
> 

I want to start with Len's idea for ALT and LONGDESC on 
table cells.

In Scott Luebking's experience, table cells were comprehensible
if the headers that govern a table cell were presented with the
cell content.

I think that this is the key requirement for long list-outs
of tables as well as for interactive cell-by-cell navigation.
We in the HC team, I would estimate, have been operating on the
tacet assumption that it will be easier to get authors and 
author tools to trace these linkages to where the "documentation"
of a table entry appears in the table than to get them to write
extra documentation in ALT and LONGDESC areas.

Further, as Jason pointed out, with the proposed broader usability
of LINK and META one can put 

	the functional equivalent of an ALT in a META on a table cell
	the functional equivalent of a LONGDESC on a table cell
		via a LINK element.

That is the general plan: first use the information that the
author has placed in view in the table cells better by making
the connections clear with SCOPE (for simple cases) and AXES
(for the exceptions).

The name of the AXES attribute is a relatively smaller matter,
even if it is a real burr under _my_ saddle.  I have objected to
the name "AXES" because as a recovering mathematician, I have
expectations that go with the "axis in multidimensional space"
concept that this attribute, when you figure out how it should
really be used, doesn't comply with.

To me, HEADERS would work.  We can tell people to use this for
both nominal and de-facto headers: what you would need to include
in "a one-sentence explanation of what a table cell tells you."
This may mean it points to things that are TD cells in their
syntax; but I don't think the authors will be much confused by
that.

Which is a roundabout way to say:

"HEADERS" captures what one should do with this attribute better
than "AXES."  "HEADERS" works for me.

-- Al Gilman

Received on Thursday, 23 October 1997 08:54:00 UTC