Re: rationale for standardizing "start-reading" bookmark name

to follow up on what Chris Maden said:

> [Al Gilman]
> > The basic reason for standardizing the name of the target anchor is
> > that there are two people who have to implement part of the linking
> > function involved: the better we name the function by what we tell
> > them to name the target, the better they will implement their two
> > pieces in a consistent and constructive way.
> 
> Wouldn't standardizing the *relationship* attribute value make more
> sense?  I realize that very few browsers support rel currently, but
> even fewer support the standardized skip-the-headers anchor name, so
> we may as well standardize a value of the attribute intended for
> precisely this purpose: expressing the relationship between ends of a
> link.

This proposal does not wait for browsers to adopt any
skip-the-headers new function.  It would work in legacy browsers
back before HTML2.  The author does that with a link from the
head of the header to the start-reading point, which the audio
visitor has the option to activate.

In addition, these two attributes don't fill the same function.
The REL attribute on an actual link, in its highest and best use,
describes the relationship of the target as seen from the origin
of the link, when all these things are known.  But the TITLE of
the link anchor is where this information belongs in non-standard
natural language.  Only if the person browsing the referring page
wanted something automatic done with links of a specific REL
would there be a point in standardizing this.

But in general this is not automatic.  When a TITLE or REL is set
in the referring anchor, the target page and the referring page
exist and the REL can respond to the specific relationship that
they enjoy.  The NAME attribute on a bookmark, on the other hand,
describes the purpose of the jump from the perspective of the
site receiving the visitor, and the most important customer to
reach is an unknown author of an as-yet-unwritten page.

Al

Received on Thursday, 5 November 1998 21:14:45 UTC