Re: Cleaning House

Jeff Cutsinger wrote:

 > Show me an example of a popular WYSIWYG editor that does this.

We're designing an HTML dialect for tomorrow, not for today.
Forget what current editors do : think about what tomorrow's
editors /could/ and /should/ do.

 > a) Using a presentational paradigm isn't treating users as idiots. To
 > users, a computer is a tool. They don't care about "semantics v.
 > presentational markup", they care about getting that timeliness report
 > done yesterday and looking great for their boss. I'm thinking of some
 > (very smart but computer illiterate) people at my office who have a hard
 > enough time getting the conceptually simpler presentational paradigm down.

It's /really/ not difficult to explain to someone that a stretch
of text is to be emboldened /for a reason/.  No-one emboldens for
fun.  All that is necessary is to change their mind-set.  Then
show them how easy it is to change all instances of (say) keyword
from bold black to boxed red while leaving all other emboldened
text unchanged.  They'll never look back.

 > b) Your example works fine until the semantic markup gets embedded into
 > a document with a different stylesheet that is out of their control.

Then their markup will be rendered according to the wishes of
the document author, as is correct.  There is undoubtedly
a lacuna if the element set of the source document is not
a subset of the element set of the destination document, and
this is an interesting issue that merits further research,
but the underlying theme of providing the user with /only/
semantic markup is (IMO) surely the only logical way to go.

 > c) Regardless of your point, there are lots of WYSIWYG editors out there
 > that have "b" buttons. Whether this is right or not is moot: they exist
 > and will continue to exist for a long time, and markup should be able to
 > handle this.

None of these editors emit HTML 5; they all emit tag-soup, usually
claiming to be HTML 4.01.  Some of the better ones (Dreamweaver,
for example, and the late-lamented HoTMetaL PRO) actually /do/
emit real HTML 4.01, at least until the user causes the editor
to bypass its internal syntax checking.  But our challenge is
to define HTML 5; leave the WYSIWYG editor writers to worry
their side of the problem and don't conflate only marginally
related issues.

Philip Taylor

Received on Thursday, 3 May 2007 14:10:25 UTC