W3C home > Mailing lists > Public > www-style@w3.org > September 2004

Re: Problem with 'target-new' property and alternative suggestion

From: David Woolley <david@djwhome.demon.co.uk>
Date: Mon, 13 Sep 2004 22:31:49 +0100 (BST)
Message-Id: <200409132131.i8DLVox00853@djwhome.demon.co.uk>
To: www-style@w3.org

> A browser cannot know whether gmail.google.com and google.com are 
> "external" to each other, and they are on the same domain. It's not 
> uncommon to have https:// and http:// sites "internal" to eachother 

In particular, I think "internal" is used here to represent
the condition under which authors think that readers should be
forced to have a new window.  Charitably, that could be defined as
meaning "NotInMyEditorialControl" and more realistically, although
not something the average author would want to make explicit,
"LessLikelyToResultInUsMakingMoney".  Generally the purpose of
forcing a new window is to try and lock people into the site; in other
words it is not selfless indication of a semantic difference but a
selfish attempt to gain a commercial advantage.

Note, also, that the boundary between internal and external can occur
within a single domain name (domain names are primarily physical
navigation properties, although larger businesses try and ensure that
they correspond to business unit boundaries). 

> also, but most browser would consider them separate "websites".  I 
> personally like the rel="external" idea. I even imagine further useful 
> relationships, such as "help" and "sidebar" which could significantly 

In reality, rel=sidebar more or less exists as rel=contents.  In my
view, if graphical browser developers had developed them as information
consumer tools, they might have made heavy use of link and made things
like frames unnecessary for consumers.  The money, however, is with
information providers, and those providing the money have not bought
into the semantic web concept, so the browsers have really been 
graphical rendering engines, and don't do imaginative things with
semantic information (designers would not want the uniformity of user
interface that would result, because it would limit there ability to
brand the site).

(For example, a deep link to a body text page could have automatically
caused the browser to display the immediately superior navigation 
frame because of a link rel=contents.)

> reduce dependence on script to open undecorated windows. And these "rel" 
> values are not incompatible with a CSS target-new property... CSS 
> selectors could even be part of the UA stylesheet to implement these 
> "rel" values.

CSS target=new is certainly compatible, if you take the position that
CSS includes behaviours, but HTML target is, primarily, an attempt to
specify concrete behaviours, and doesn't really belong in a semantic
markup language.
Received on Monday, 13 September 2004 21:32:18 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:32 GMT