Hi Ulrich,
Unfortunately I have not yet found enough time to finish support for
this, and my experimental branch is horribly out of date. However, the
basic idea is as follows:
For Internet Explorer little work is needed because VML elements are
selectable. In fact it works even in the current version, but only one
word at a time (you can't select individual letters). Note: this
changed in IE8, but as IE8's VML implementation is still quite buggy
we'll just have to wait and see.
For WebKit-based browsers there is -webkit-user-select, which can be
used to achieve the illusion of proper text selection.
For Gecko there's the similar -moz-user-select.
For other browsers you can usually use a transparent image which the
user can then select. Unfortunately this does not work in Opera.
In all cases, the user actually ends up copying the (hidden)
alternative text that is meant for printing. So the user doesn't
actually select the text itself, but another element that looks like a
selection. It works remarkably well, but it doesn't come for free
either from a performance point of view. So I've been thinking about
limiting text selection to words only which would not be so bad.
Anyway, this IS coming some day :)
Simo