Webmaster Forum / HTML, CSS, Scripts / CSS / December 2008

But, with respect, that is not me saying that a table is a list in some
general unconditional way. If some As are Bs, it does not follow that
all Bs are As.

Many tables, for one thing, are composed of many lists and the point of
these lists being in the table format is to reveal relationships between
various list items across the lists. Thus you might see a list of types
of things: animal, mineral and vegetable in one row or column; cats,
salt and carrots in another row or column. You might see a list of
numbers in one row or column and a list of other items in another row or
column.

A limiting case is a table with just one row or column. A one row or a
one col table is a legal HTML entity. And, I would say, is not the worst
thing you can use to display an unordered list if, for any reason
whatsoever, you could not make do with the particularly specialised 'one
note' UL element.

Generally, for unordered lists, you would have to be very fussy indeed
not to be able to make do with a UL. But, contrary to what we might call
the Naive Theory of HTML Semantic Structure (NTHSS), a table with a
suitable table heading like: "Items to be bought". Under which are the
cells with boring stuff like "low fat cheese, skim milk, eggs..." would
certainly *seem* to do fine too.

Now, the NTHSS would say not only that the right semantic element to use
for an unordered list is a UL but that a table is a *semantically wrong*
thing to use. And it would say this in the face of clear evidence that
anyone seeing a table as described just above would be in no doubt
whatsoever about its meaning. Either a UL or a Table (in suitable
context or table headed) would be about as semantically equivalent as
any two things could be in the normal meaning of such a phrase.

I pass now from this matter with unordered lists to ordered lists. An
ordered list is a series of items where the order matters. The order
needs to be reflected in some way. This is not something that an author
can ignore. If an OL is used without the numbers (they can be styled
off) and the context does not make it clear that it is an ordered list,
then there is no semantic structure involved worth a pinch of salt.

The OL, in and of itself, is nothing much at all. It is really, denied
of course, by the NTHSS, a mere hand maiden to presentation, a means to
presentation. Some basic presentation is the heart of communicating
meaning to humans. It is not the other way around as many people who
hold the NTHSS theory maintain.

So, as is sensible, an author will mostly leave the numbers in to help
show that the list is ordered. (*help*, because the mere presence of
numbers is possible with unordered lists, the numbers in an OL could be
misinterpreted as mere markers for reference). But suppose he is not
happy with how the numbers line up to the list items. He is fussy! He
has aesthetic barrows to push. Good for him!

Suppose he can achieve the fussier look he wants by using a table. Well,
what is wrong with this? Nothing, on the semantic front. If the numbers,
which are a list in themselves, are put into a left column, and the
things to be done (in some algorithmic example) are listed in cells in
the adjoining column, and the author shows the tabular relationships row
by row, then no human will be at a loss about the meaning. There are
these two ways to communicate the ordered list to a human. That is what
I mean by semantic equivalence. That is what it is in effect.

There is no hidden deep meaning that attaches to elements that sort of
of float next to them. Almost the entire function of elements is to be a
good and efficient vehicle for presentation. If either of two elements
deliver the same ultimate look that humans will *understand* in the way
the author intends, then we have semantic equivalence. This look is a
powerful means of communicating something to visual users. Similar
remarks, in respects that are relevant, can be made about aural
presentation.