Re: Comments on HTMl 4 draft (9/Nov/1997)

[Re <I> and suchlike]

   > Were I to use the tagline you write above, I would have
   > enclosed it in <A HREF...>, effectively 'setting it apart'
   > from the rest of the paragraph, anyway.  Have you another
   > example? 

Why in goddess' name are we arguing about stuff that was thrashed to
death five years ago and which is heavily documented all over the

   EM and STRONG are for marking up emphasized text 
   CITE is for (usually italicized) citations
   KBD, CODE, SAMP, and a bunch of others for are (usually typewriter) code
   I and B are for italics and bold to make things look pretty (ie when 
     there is no other element to do the job, and you are writing for
     users with no access to a styling enging). Same goes for FONT.

Making text look pretty is part of the job of a graphic designer.
There is nothing wrong with making text look pretty, provided it
doesn't interfere with the primary purpose of text, which is to pass
information from the author to the reader.

Making text do this effectively is part of the job of a text encoder
or doc engineer.  Most authors are neither encoders nor doc engineers
nor designers, but they still feel qualified to make a shot at it, and
it should not surprise anyone that most of them fail.

HTML, even with CSS, is severely limited as a markup language because
it describes one type of document only. XML lets you describe any type
of document you want, and 1998 is likely to provide more software for
XML and XSL that you'll know what to do with. 

I haven't authored anything directly in HTML since 1994 and I have no
intention of encouraging others to do so, except for the most trivial
purposes. Authoring in "full" SGML or in XML means you can convert to
any kind of HTML (using something like OmniMark LE) to get your stuff
onto the Web until such time as XML general-purpose browsers are
widely used. You can hand the resulting HTML to a designer to prettify
if like me your graphic design skills are not your strong point :-)

If you prefer to use FrontPage/NSGoldEd or other graphical front ends
to HTML for origination, you must understand that what you generate is
not conformant HTML (with one or two exceptions like HoTMetaL) and
cannot reasonably be converted to XML by any meaningful automated
process unless you have very large volumes and are prepared to write
the heuristics into your own conversion system: for most users it is
going to be a long, slow, manual editing job.

All of which makes a discussion of <I> and <B> vs <EM> and <STRONG>
a little redundant.


Received on Friday, 26 December 1997 12:21:17 UTC