Re: Ems in tables

David - Morris (dwm@shell.portal.com)
Wed, 29 Mar 95 04:34:07 EST

On Sun, 26 Mar 1995, Terry Allen wrote:

> Folks, there's a lot of experience with rendering SGML tables of the sort
> proposed here. Please listen to it.

Well I have lots of experience using tables with IBM's GML and I am
convinced that it is not a sufficient solution to restrict HTML to
proportional sizes. Fixed measurements don't make much sense since the
browsers are not going to find it easy to learn the meaning of an inch
(or mm or pica etc.) on the user's display. ems have value that the
in some sense correspond to what the user has found to be a readible font
size. Surely we can step above the debate about how hard it is to know what
the ems is when the table begins or when the table contains multiple fonts
and find a simple definition. What matters here is for the author to have
a way to specify the width of the table in units that are meaningful
in the ultimate consumer's context such that the information intended
to be conveyed by spatial relationships is preserved. Seems pretty
reasonable to have the browser and user agree on the notion of a
normal font size and render all table ems on that basis. Browsers could
provide a table default font which differed. I find it hard to believe that
every browser won't know what font size it would use for the table if no
other size is specified/implied. Use that value if you prefer.

Dave Morris