I seem to have stirred up some controversy with SGML Open by saying I like ems
in tables as well as for general measurements. I still stand by that position,
though I also accept Paul Grosso's observation that ems can easily cause
extremely degenerate tables if many point sizes are mixed, particularly if
there are wide divergences (like 9 point and 14 point in the same table).
Rather than have tables beccome unmanagable, I'd eliminate ems there, but if
we're going to allow users to create typographic specifications, we need ems
elsewhere.
As I told Paul, I think allowing typographic specifications in HTML is a VERY
BAD THING. Any information providers who are crazy enough to mix 9 point and
14 point in a table (unless the 14 point were an isolated integral sign)
deserves to have their documents unread. As I said in previous postings, it
make sense to have generic identification of emphasis and the like (we can't,
after all, do the sort of thing I do in my local documents and have
information-identification tags for all sorts of things like biological names,
book titles, etc., and still make HTML an application that many people will
willingly adopt). But we need to accept a new model of publishing that allows
detailed typographic specification to be shifted away from the generators to
the consumers of information. Let the consumers decide how they want titles,
emphasis, and the like to look. Get the providers out of thinking of HTML as
another WordPerfect. Power to the consumers!!
(A lot of the signature lines don't come across on my mail system, so I don't
know who at EBT made this posting to which I am replying. I'd appreciate if
the poster would contact me.)
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 @ ornl.gov