Re: Anchor dilemma

Foteos Macrides (MACRIDES@sci.wfbr.edu)
Mon, 19 Feb 1996 14:30:42 -0500 (EST)


Date: Mon, 19 Feb 1996 14:30:42 -0500 (EST)
From: Foteos Macrides <MACRIDES@sci.wfbr.edu>
Subject: Re: Anchor dilemma
To: william@cs.columbia.edu
Cc: www-html@w3.org
Message-Id: <01I1EF3ETRV0000YZW@SCI.WFBR.EDU>

"William C. Cheng" <william@cs.columbia.edu> wrote:
>You wrote:
> | "Daniel W. Connolly" <connolly@beach.w3.org> wrote:
> | >In message <01I1E1S7LDBM000V2A@SCI.WFBR.EDU>, Foteos Macrides writes:
> | >>"Daniel W. Connolly" <connolly@beach.w3.org> wrote:
> | >>>This "empty anchors" but is annoying, as it causes lots of folks
>
>Dan's typo here==>     ... bug ...
>
> | >>>to write:
> | >>>
> | >>>	<h1><a name="target"> blah blah </h1>
> | >>>
> | >>>without closing the <a>, because their generator software doesn't know
> | >>>where the end of the target is.
> | >>
> | >>	Has <SPOT ID="target">, which solved the empty anchor problem,
> | >>been abandoned?
> | >
> | >Putting <spot> in a spec won't fix the empty anchor problem: the
> | >implementations have to be revised. I'm not interested in using <spot>
> | >to band-aid the problems with <A>. I want the bugs fixed.
        ||||||||
	^^^^^^^^

> | >
> | >I _am_ interested in using <spot> for other things (like change bars,
> | >and other non-hierarchical structures). So in my mind, it hasn't been
> | >abandoned. I can't say it's high on the agenda right now, though.
> | 
> | 	Perhaps I haven't grapsed what is the problem with <A>.
>
>The problem is that <A NAME="target"></A> is legal in HTML-2.0 but some
>browsers flag it as error and don't handle it properly.  These browsers
>insist that there must be some text between <A ...> and </A>.
>
> | People with clients which did not implement SPOT are tempted to use:
> | 
> | 	<A NAME="target></A>Blah blah ...
> | 
> | for what:
> | 
> | 	<SPOT ID="target">Blah blah ...
> | 	
> | does.  Using:
> | 
> | 	<A NAME="target">Blah</A> blah ...
> | 
> | avoids the problem of an empty anchor, ...
>
>The bugs are in the browsers, not in an HTML document.
>
> | small portion of the text as content is not logically what's intended.
> | Also, it achieves the desired result only for clients which have stopped
> | highlighting anchor content if the anchor tag lacks an HREF attribute.
> | In those cases, forgetting to include an end tag for the anchor has
> | no visible consequences (the client doesn't highlight, or otherwise
> | indicate a link, for a large portion or everything which follows the
> | HREF-less NAMEd anchor start tag).  So naive users of such clients
> | draw the incorrect inference that an end tag is not required for
             |||||||||||||||||||
	     ^^^^^^^^^^^^^^^^^^^

> | NAMEed anchors without an HREF.  That's not a bug in the clients which
> | needs to be corrected, but is a predictable consequece of the clients
> | not supporting markup which is explicitly designed for what is intended.
>
>The HTML-2.0 DTD says:
>
>    <!ELEMENT A     - - %A.content -(A)>
>
>This means that </A> is *not* optional!  Therefore, missing </A> is
>bad HTML!

	Applying error recovery when rendering bad HTML is not a bug,
but in this case can lead the the *incorrect inference* by naive users
of the client that the end tag is optional when the anchor start tag
does not incldue an HREF attribute.

	A band-aid unlikely to heal the wound is leaving valuable
HTML 3.0 out of active W3C working drafts and I-Ds if the so-called
"major" clients thus far have ignored it (IMHO).

				Fote

=========================================================================
 Foteos Macrides            Worcester Foundation for Biomedical Research
 MACRIDES@SCI.WFBR.EDU         222 Maple Avenue, Shrewsbury, MA 01545
=========================================================================