Re: Target attrib in HTML 4.01 Strict

> However, Strict is not supposed to be used with Frameset as frames (except for
> some reason 'iframe') are deprecated themselves. XFrames has some crude URI
> syntax that doesn't require an 'target' attribute anymore AFAIR.

For HTML 4, all of these features were included in the Transitional DTD,
and therefore implicitly deprecated (Frameset is just a wrapper that
sets a conditional and then invokes Transitional).

>   <a onclick="open_new_window(this.href);return false;"
>      href="foobar.html">

The return false shouldn't be hard coded at this level as this may
cause dead links for people with popup blocking turned on.  A false
return needs to be conditional on succesful creation of the new
window.

> so for behavioural, which should rather and can already be handled by
> scripting languages.

Yes.  I agree that the objection to target="_blank" is that it is behavioural
(and also somewhat undesirable for users).  (The objection to general
targetting is the same as for frames, namely that it creates structures
that violate a fundamental principle of the web that all resources should
be capable of being specified by a URI.) 

> That's very annoying behaviour. Users who want (some) links to open in new
> windows or tabs, usually know how to achieve that. Only very few pages should
> *offer* that feature independent of the browser, for example search engine
> results (Google does, for instance).

Agreed.  However, there is a problem that commercial designers believe
that it locks people into web sites (bad for the web, believed as good
for the company by some marketing people) and some company lawyers fear
that same window links will be assumed to be the legal responsibility of
the referring company (an even more annoying features is where you cannot
directly follow a link but must first read a legal disclaimer).  There is
an argument that the alternative (as few designers know the scripting
fallbacks that you used and href="#" is the "preferred" form) of having
a large proportion of dead links on commercial sites, with scriptng off
or unavailable, is worse than, at least, having the links with undesirable
behaviour and/or unreferencable resource combinations.

(My own view is that the only sensible way of implementing parallel
tables of contents would have been for browsers to use the, long defined,
link rel="contents" to automatically bring up the relevant contents 
page (this wouldn't work with current graphical designs for pages). But
most browsers aren't really information browsers, but GUI remoting/
branding support, tools.)

Incidentally, common GUI browsers allow following links by dragging them
into another window, so, with some education, any target free site
can be made to behave as a multi-pane site by a user.

Received on Wednesday, 11 February 2004 03:49:22 UTC