Pedantic Forms Test

In the HTML 2.0 spec, section 8.1.2.4 explicitly states that:

At all times, exactly one of the radio buttons in a set is checked. If none of the INPUT elements of a set of radio buttons specifies `CHECKED', then the user agent must check the first radio button of the set initially.

Note however that at least Netscape 2.0 ignores this, so that if no radio buttons in an array are CHECKED, as in this form:

initially none of the radio buttons will be selected. I happen to think the spec is correct, and HTML authors that rely on this feature are writing bad surveys. The reason is that as currently implemented, once a radio is selected it can never be unselected unless the entire form is reset. To me, this violates the principle of least surprise.

HTML authors should, in these cases, include a None of the Above or Other radio and have this be the default checked choice, like so:

But oh well, since the other browsers do it this way, Grail now exhibits this behavior by default. It's more fun to be pedantic though, so the undocumented parsing-html--strict preference, when set, will force Grail to conform to the spec as written.

Grail now exhibits the following behavior, which is identical to what Netscape 2.0 exhibits:

Initially no radios in the array are selected
Once a radio has been selected, there's no way to just unselected it, short of resetting the entire form.
Upon `reload' of the page, any selected radio stays selected. reload does *not* reset the radio. Note that Microsoft Internet Explorer 2.0 does reset the radio to none-selected upon page reload. So who's broken behavior should Grail mimic, hmm?
Using the form's reset button, all radios are unselected.