[UPHPU] Validation

[Mac Newbold]

Today at 12:28pm, Connor Boyack said:

> On 6/6/06, Mac Newbold <mac at macnewbold.com> wrote:
>> 
>> I think I might disagree with you on this, by saying that layout is one of
>> the perfectly valid and reasonable purposes of tables. In fact, I could
>> say that marking up tabular, spreadsheet-like data was the exact purpose
>> of a table, so they are _purely_ for layout, but I don't know if I can go
>> that far. In any case, I may just have to agree to disagree on this point.
>> 
> Layout is horrendous for tables... In days gone by, it was an accepted and
> common way of doing a layout.  But all it really does it add bloat to your
> markup, and restrict you from making any design changes.  Hence the
> popularity and adaptability of sites featured at the CSS Zen Garden - using
> the same markup, you can have completely different layouts.  Here at Utah
> Interactive we just released the new design for Utah.gov - a re-design that
> we saved countless hours on, because the previous one was CSS-based, so
> there was very little tweaking and coding.  CSS is a time saver, both for
> development, and end-user download.  Tables (for layout), spacer GIFs, and
> tag soup should be flushed down a toilet.

I agree with you that there are sites, probably even many sites, and maybe 
a majority of sites, where the method you describe works much better than 
tables. But for fluid layouts, especially ones in which could be 
considered "tabular" with at least two rows and two columns that need to 
match up yet be fluid and dynamic in their width or height, going 
table-less is generally impossible in my experience. With divs alone, you 
can get a row to line up right, or a column to line up right, by nesting 
the "cells" inside that div, but you can't nest a "cell" inside a div for 
both the column and the row at the same time, so you can't maintain both 
the proper height and the proper width at the same time unless you make it 
static and non-fluid.

I'm not trying to advocate that tables are the only way to go. I don't 
think anyone rational could honestly push that position any more. I'm just 
trying to say that _never_ using tables is not any more correct than 
_always_ using only tables. Most of us seem to agree on that, we just 
disagree on what uses of tables are "acceptable" and which are not. I 
think there are more good uses of tables than a lot of others would agree 
with. I also have a different bias than many people participating in this 
dicussion: I'm about the bottom line, getting it done quickly, 
maintainably, working cross-browser, and designed with an eye to the 
future, regardless of whether or not the methods used "validate" or fit 
within the definition du jour of "Good"(tm), except where the two aspects 
happen to coincide.

I also would like to advocate fixing the standards where necessary, or 
fising the definition of "Good" where necessary, so that there is a "Good" 
way to make fluid dynamic layouts of rows and columns, and a good way to 
do vertical-align:middle without using an unnecessary table. Hopefully 
some day someone who has a measure of influence will do something about 
these issues, and the browsers will implement the new standards, and 
everyone will be happy until we find another pseudo-religious battle to 
debate.

Mac

--
Mac Newbold		MNE - Mac Newbold Enterprises, LLC
mac at macnewbold.com	http://www.macnewbold.com/