Re: missing principle

On 4/28/07, Smylers <Smylers@stripey.com> wrote:
>
>
> Maciej Stachowiak writes:
>
> > On Apr 28, 2007, at 1:11 AM, Mike Schinkel wrote:
> >
> > > Anne van Kesteren wrote:
> > >
> > > > On Sat, 28 Apr 2007 09:44:27 +0200, Mike Schinkel <w3c-
> > > > lists@mikeschinkel.com> wrote:
> > > >
> > > > > True, but the important point is that a UA can never be sure why
> > > > > the author used it so it can't be trusted as a semantic element.
> > > >
> > > > That argument goes for *any* element.
> > >
> > > Actually, it is not true for any element.  Some elements, such as
> > > <em> are almost never used except when the user wants to emphasize.
> >
> > Lots of use of <em>  on the web appears to be to italicize for reasons
> > other than  emphasis, because authors have been taught that <em> is
> > "more  semantic" than <i> .
>
> Or because the content's author used a wysiwyg editor and clicked on the
> 'i' button, which (misguidedly) inserts an <em> element.



Ah, I just realized that <em> and <strong> actually carry less semantic
payload than <i> and <b> - in terms of real human semantics.

Back when <em> and <strong> were introduced, I followed the recommendation
for a while and used those elements in place of <i> and <b>, though even at
the time it was not clear to me that this was in fact better.
Never was able to convince myself this was indeed an improvement, and
eventually reverted back to <i> and <b>.

Finally made the missing connection while following this discussion.  When
communicating among a group, there is a shared set of knowledge, customs,
and conventions with specific meaning - within that group.  In effect
specific semantic interpretation of a typographic convention is determined
by the context (pragmatics).  Specific typographic conventions often carry
specific semantic meaning within a conversation.

Which makes <em> and <strong> exactly wrong, except in the less usual case:
conversations where typographic conventions carry no semantics.

Browsers know nothing about the human context.  A browser that chose to
represent <em> as something other than italics runs the risk of removing
semantics.