Re: Ems in tables

Joe English (
Sun, 26 Mar 95 21:07:54 EST (Yuri Rubinsky) wrote:
> Jon Bosak ( > )writes/Lee Quin (> >) writes:
> > > It seems to me that a perfectly simple and well-defined interpretation
> > > of ems in table column widths is that they refer to ems at the start
> > > of the table - i.e. the type size in use when <table...> was encountered.
> > So if the sentence immediately preceding the <table> tag is
> > ... the command to format a diskette is FORMAT A: /U.
> > (where "FORMAT A: /U" is in monospaced Courier but the rest is in
> > proportionally spaced Times) then "em" in the table that follows means
> > the em of the Courier font. But if I change my mind and edit the
> > sentence to read [...]
> > then "em" in the table that follows is now suddenly the em of the
> > Times font.
> It's actually a little worse than this: It means you can never have
> a table immediately follow an H1, H2, etc.
> Unless we decide -- which seems completely counter intuitive -- to say
> that a table's default point size is the same as a paragraph's.

This would be a more sensible interpretation.
It shouldn't matter if a table *follows* a heading,
or a <P> whose last element is a <CODE>; since the
table does not appear *inside* those elements
the "current font" should be that assigned to the
<BODY> (or <DIV> or whatever) element which directly
*contains* the <TABLE> element.

> Or follow the advice of so many suggestions and give up on ems. Do the
> proportioning after the table is calculated.

This would be even more sensible.

Who wants to count ems when making a table anyway?
Proportional widths are easier to work with for both
browser and author.