Terry Allen <[log in to unmask]> wrote:
|[rendering info in start tags] *is* the wrong thing, as it completely subverts
the principle of structured markup and the semantics of the DTD, which, as we
see here weekly, are important to many users and vendors.
So long as the document is otherwise conforming and all element-level style
info in crammed into a single attribute, how is this true?
|And as Tim Bray remarked, it totally screws up any downstream processing of
the document, which is important to the Internet as a whole.
|Tim Bray ([log in to unmask]) Fri, 2 Feb 96 08:12 PST
|> <P FONT=MT-GillSans-BoldCondensed SIZE=18>James Chalmers Arrives</P>
|is a bomb-crater, a disaster, a severe information loss, but *any* of
|> <H2>James Chalmers Arrives</H2>
|> <H2 CLASS=leader>...
|> <H2 STYLE=L15>
|are much better.
Tim's first example, the bomb crater (a widely distributed bomb crater at
that), is the impetus for incorporating something along the lines of <H2
STYLE=L15> or perhaps <H2 STYLE="css: font-family: teletype, font-size: 18">
into the draft-on-the-table.
If specifying style were permitted by <LINK REL=STYLE>, by <STYLE> or by <ELT
STYLE="..."> and the document is otherwise conformant HTML, how does the
author's choice of style mechanism effect document interchange?