Ian Hickson wrote:
HTML5 requires that there not be a reload. Setting location.hash
eventually (if you follow the admittedly convoluted definitions) is
equivalent to running the navigation algorithm:
http://www.whatwg.org/specs/web-apps/current-work/#navigate
...which, in step 4, just scrolls and aborts the algorithm without
actually changing the active Document object.
Hmm. That algorithm doesn't look quite right in the case when going
from http://example.com/#foo to http://example.com/. That case should
actually do the load instead of scrolling, last I checked, unless done
as part of a history traversal. See also
<https://bugzilla.mozilla.org/show_bug.cgi?id=273158>.
Same thing for the case when going from http://example.com/ to
http://example.com/ (as comment 5 in the above bug points out). If I
read the current draft right, per step 4 that would be a no-op, whereas
in practice I don't believe it is in browsers...
-Boris