[whatwg] Target Attribute Values

[Smylers]

Spartanicus writes:

> Smylers <Smylers at stripey.com> wrote:
> 
> > But _requiring_ user agents to offer opt-outs seems excessive, and
> > possibly beyond the jurisdiction of the spec.
> 
> Possibly, but then what's the point of making _blank non conforming if
> it is not trying to be a benefit to users by discouraging its use.

Well that's begging the question: Ian has (tentatively) made it
non-conforming while saying there are arguments for not doing.

There's also a difference between marking something as non-conforming
(because there's a better alternative which should be used instead), and
completely blocking the old way of doing it.

For example, we could agree that it would benefit users by authors no
longer using the <i> element (instead picking <em> or <dfn> or <cite> or
whatever is appropriate); but it wouldn't benefit users for browsers to
completely ignore <i> tags, rendering its contents in Roman such that
users can't discern there was any mark-up there.

If target="_blank" is ignored users can't tell that the author intended
some behaviour there.  Or perhaps a help link has target="_blank" and is
labelled with "opens in new window" -- which could be dangerous if a
user believed that label, only to lose her partly completed form.

Smylers