Moving beyond ems in tables (Was: Re: Ems in tables)

James D Mason (
Mon, 27 Mar 95 10:03:37 EST

Since I'm one of the ones who kept this thread going (by suggesting we might
still want ems in tables), I'll say that we've seen enough discussion to say
it's not a good idea to keep ems in a totally uncontrolled environment.

Most of us who, like Jon, Paul, Terry, Yuri, and myself, come out of
traditional (often technical) publishing environments have a pretty good idea
of what we'll do with tables and so know what risks--or lack of risks--we run
in our selection of ways to manage column widths. On the Web, who knows what
tables will be used for? (I'm thinking of tables for something like drama
scripts myself because there isn't much other way to render a scheme involving
actions and responsible parties.) So ems may be dangerous.

In saying that we ought to get past the ems-in-tables issue, though, I can't
help remarking that it's brought us back to the issue of what can happen where
in an HTML document. This time we're asking in what circumstances a table can
appear. A couple of months ago the issue was what can happen inside a table.
As I said back then, I think sooner or later the HTML community is going to
have to face up to the issues of containment and hierarchy, even if it means
an envirironment in which two or more DTDs coexist in parallel. Whether it's a
matter of a wide-open but simple DTD along with one that is rich but
structured or a matter of a DTD for people who want to author in HTML and
another for those (like Jon and me) want to use our own DTDs and and let a
Web server translate into some form of HTML for distribution remains to be
seen. But we need to think about it.


Dr. James D. Mason
(ISO/IEC JTC1/SC18/WG8 Convenor)
Oak Ridge National Laboratory
Information Management Services
Bldg. 2506, M.S. 6302, P.O. Box 2008
Oak Ridge, TN 37831-6302 U.S.A.
Telephone: +1 615 574-6973
Facsimile: + 1 615 574-6983
Network: masonjd @