RE: Technique 3.4.1 Check document for relative units of measure

I like to think there aren't cases where an absolute size is needed. A quasi
exception is things like image width and height that accept only pixel
measurements (appropriately given the nature of images). Especially with
tables, I use percentage sizes if I want it to scale, and no sizes if I want
it to stay as small as it can on the basis of its contents. I have been
known to sneak in a transparent image if I wanted to impose at least a
minimum size on a table cell though...

On forms - I think if a table is being used to lay out a form, the scaling
of the table shouldn't affect layout aside from adding whitespace. The
author does need to be aware of potential font size scaling impacts on the
form layout, so if the table can accommdate that, it shouldn't need a size
constraint.

Frames is tricky - there certainly can be cases where it seems there is an
appropriate fixed width for a frame. But I think Charles is right that that
fixed width may not always work. I usually use a relative width that works
out to be close to the preferred width on most monitors, then make the frame
resizable so the user can correct if needed.

Michael

-----Original Message-----
From: w3c-wai-er-ig-request@w3.org
[mailto:w3c-wai-er-ig-request@w3.org]On Behalf Of Chris Ridpath
Sent: Thursday, September 28, 2000 12:00 PM
To: Michael Cooper; WAI ER IG List
Subject: Re: Technique 3.4.1 Check document for relative units of
measure


Could there be some cases where an absolute size in tables and framesets may
be appropriate?

An absolute size is OK if:

If the table contains a form - the author may not want the form layout to be
changed by a table changing size.
If the table column or frame contains images (navigation buttons for
example) that are a set size - the other table columns or frames may change
size but the column containing the images should stay put.
If the entire table or frameset size is less than 640 X 480 pixels - this
will fit on the lowest resolution monitor.

Chris


----- Original Message -----
From: "Michael Cooper" <mcooper@cast.org>
To: "WAI ER IG List" <w3c-wai-er-ig@w3.org>
Sent: Thursday, September 21, 2000 10:38 AM
Subject: RE: Technique 3.4.1 Check document for relative units of measure


> For Bobby, I went through the HTML spec and we've only implemented the
> requirement to have relative size units for those elements that support
> them. When we originally had it for any size attribute, every table
border,
> cellspacing, cellpadding, etc. attribute got called out, even width and
> height for images!. But in HTML, there are no relative sizes you can
define
> for those attributes. In CSS, there are (and by the way, you can use a
> fractional em, like "border: .1em"), so if we were evaluating CSS I would
> say border, padding, margin etc. should be covered.
>
> The only elements and attributes we check for absolute size, then, are:
>
> COL - width, charoff
> COLGROUP - width, charoff
> HR - width
> FRAMESET - rows, cols
> IFRAME - width, height
> TABLE - width
> TBODY - charoff
> TH - width, height, charoff
> TFOOT - charoff
> THEAD - charoff
> TD - width, height, charoff
> TR - height, charoff
>
> Michael
>
> -----Original Message-----
> From: w3c-wai-er-ig-request@w3.org
> [mailto:w3c-wai-er-ig-request@w3.org]On Behalf Of Chris Ridpath
> Sent: Wednesday, September 20, 2000 12:07 PM
> To: WAI ER IG List
> Subject: Technique 3.4.1 Check document for relative units of measure
>
>
> Could the 'border' attribute be an exception to this rule? It's a common
> practice to use 1 or 2 for a table/image/frame border to indicate that
there
> should be some sort of thin line surrounding the object. If we do require
a
> relative measure for a border, what would it be? (I think that an 'em' or
> 'ex' would be too large to replace a 1 pixel border.)
>
> Chris
>
>

Received on Monday, 2 October 2000 10:01:43 UTC