Re: Empty anchor

John Franks (john@math.nwu.edu)
Tue, 13 Sep 1994 09:56:49 -0500 (CDT)


According to Bert Bos:
> |
> |1. Is the empty anchor legal HTML?
>
> It is. ("empty" in the sense you meant, not in the SGML sense.)
>
> |2. If so is it a bug in Mosaic or libwww that causes it to fail?
> | Will it be fixed?
>
> Depends on what you mean by "fail". I haven't seen it crash.
>

I don't mean fail in the sense of crash, I mean fail in the sense
of fail to function in the correct manner. If an HTML document
contains
<a name="marker"></a>
and that that document is requested by XMosaic (and other browsers
as well) with the URL http://host/file#marker then the browser
will retrieve the document but will not position the text window
to the correct place (i.e. someplace near the "marker" anchor.

> |3. Is <a name="something"> alone legal, i.e. without the </a> ?
>
> No, it's not.
>

On the other hand, if the document contains only <a name="marker">
with no close to the tag then browsers handle it in the desired way.
My point was that an illegal HTML construct functions while the legal
construct to achieve the same effect fails. This provides a strong
incentive for document creators to write illegal HTML. The fact that
the situation will be better with HTML 3.0 doesn't mitigate the
problem much. Much of this bad HTML will still be around when HTML
3.0 browsers are common. Those browsers will either have to handle an
illegal construct or fail to provide the behavior which older browsers
did.

>
> That's why hyperlink targets have been changed in HTML 3.0. The <A>
> tag is now used solely as a hyperlink source, while almost any other
> element can function as target. For example: <a
> href="#something">... </a> can link to <em id="something">... </em>
> but also to <dl id="something">... </dl>.
>

There is still a need for a completely empty target. Can one link
to a comment, for example?

John Franks Dept of Math. Northwestern University
john@math.nwu.edu