Re: Working Group Decision on ISSUE-155 table-border

On Wed, Apr 13, 2011 at 8:01 PM, Maciej Stachowiak <mjs@apple.com> wrote:
>  If this is a generic problem (that is, if tables in non-CSS UAs
>  should basically always have borders), then tying border-drawing to
>  an optional attribute which is *not* present on the vast majority of
>  legacy tables is completely the wrong solution.
>
> Only two solutions were proposed.  We only consider proposals which
> actually are put forward.  Furthermore, no substantiation is provided
> for the assertion that this is "completely the wrong solution".

Note, this was not an attempt to suggest another solution, it was an
objection against the <table border> CP.

I thought that the above paragraph was sufficient explanation for my
assertion, but I guess not.  What I meant was that there are two
possibilities:

1) Having a border is just a nice stylistic touch, or
2) Having a border is a basic requirement for table readability.

If (1), then the border is just styling, and we don't need to worry
about providing non-CSS UAs with a styling hint, just like we don't
consider <font> necessary.

If (2), then the border is vital and should be the default styling.  I
suppose that I could have made a CP to this effect, but I didn't think
it was really in the purview of HTML, so I didn't.  This would just be
a UA formatting decision, after all, similar to how mobile browsers
sometimes adjust the width of paragraphs so they fit on the phone
screen without horizontal scrolling.

The current situation, where no-border is the default and authors can
specify a styling hint, is the worst of both worlds.  It pollutes your
markup for no good reason if (1) is true, and it means that the vast
majority of tables are still visually inaccessible if (2) is true.

This last statement is supported by the data at
<http://philip.html5.org/data/table-border-pages.txt>.  (I'm not sure
if this is new information or not.)  The vast majority of pages that
include @border set the value to 0 (about 90%, it seems), which means
that they don't have a border at all.  I don't have data on how many
tables are authored without @border at all, but I imagine it's another
substantial fraction.  So, even if border=1 is conforming, that won't
help 90+% of pages.

So, to review, if table borders are purely stylistic, we don't need to
make @border=1 conforming - it's just one of a multitude of ways in
which HTML was unnecessarily concerned with style, and most of those
are non-conforming for good reason.  If table borders are vital for
understanding the page, then making @border=1 conforming will have no
effect on more than 90% of all pages.  So, making border=1 conforming
is either silly or useless.

~TJ

Received on Friday, 15 April 2011 18:31:09 UTC