Posted by dorayme on 06/03/06 09:52

In article <96fbc$44810eb8$40cba7bd$2035@NAXS.COM>,
"Jonathan N. Little" <lws4art@centralva.net> wrote:

> Joe wrote:
>
> > No such thing as a single-column table. That's why my def says
> > "headingS".
>
> Ah but a list can have a heading, but not be a table
>
Indeed. You are getting the hang of it... there's always
something or other wrong with a definition in this area given the
background requirements at alt.html.
>
> I still say what makes a table is data arrange on a grid where the x|y
> relationship has meaning and not arbitrary positioning which defines a
> table.
>

And someone could say this can apply to a table which you would
very likely regard as an abuse.

Take

> Widgets
> --+-----------
> 1 | $10.00
> --+ ----------
> 2 | $19.00
> --+-----------
> 5 | $45.00
>

cf.

My Pics
--+------------------------
tiny | link1, link2...
--+ -----------------------
medium | link3, link4...
--+-------------------------
huge | link5, link6...

Perhaps you would concede this is a table (even if you could not
resist saying that it is not the best way of laying out this
information).

Someone would then probe you on the limits of your idea, in the
case of an html table of not more than 2 cols and one row, left
for nav, right for content, we have a relationship in a number of
ways:

To explain this, I need you to ask yourself how you are counting
tables? Simple, you say, here is a page, there is a table on it.
That is one table. Fine, but this is superficial. There is a
higher object, a table-type, if you will, instances of it on
separate pages. In the workings of a website this may well be a
far more important object.

For a start, the left content might have a certain characteristic
when it is made quite clear that one of the "links" is related to
the right content.

The table-type changes from page to page, each "link" displaying
"its content". In other words, you need to take account of a
dynamic situation with basically the "same" table on each page,
the left changing ever so slightly in content (you know the sort
of thing, a link goes deadums or is highlighted to indicate
arrival). So _this_ higher table does have a relationship between
its left and its right. You might complain that the individual
tables on individual pages fail your criterion but that is
superficial, the real goings on are to do with the way the
table-instances relate to the table-type across many pages. I am
saying that the dynamic of this higher table begins to satisfy a
similar criterion to yours except that instead of the
relationships going on between rows and columns it goes on
between these and table-types on separate pages. You will be
hating that your own criterion could be used to justify tables
for what I am certain you would regard as layout.

Now you might not like this idea of using tables for this sort of
layout but you are really pushing uphill to deny the
relationships that count. It is just that they count in a less
superficial and trivial way over a _number_ of pages.

Some tables are used for what you would say is layout but they
are ultimately, in the dynamic across pages, more deeply
something to do with exhibiting relationships between cols
across. If you want to attack this use, no good messing about
with definitions...

Perhaps it is because many folk deep down sense that what I am
saying is correct that they are skeptical of the slogan waving
that goes on in this area, the extreme fundamentalisms. They see
sense (I mean sense) in some table layouts which high priests
miss because they are looking at the wrong object.

This is not meant as a recommendation of the use of tables. And
certainly not the mindless use of them that earthlings as well as
machines spew out. It is just to say, it is pretty hard to polish
up a little definition of the situation and walk off with the
cream.

My own view is that the lines between layout and semantics are
not as clear cut as some suppose. But enough already.

--
dorayme

[Back to original message]