The behavior of the history dropdown in the location bar in 0.9.1 (tried on Linux and FreeBSD) seems to be sub-optimal. It has an entry for every URL viewed on a site, and it doesn't seem to have any reasonable sorting order. For example, say I'm using a web application on http://webapp.example.com/. In the course of using it, I hit 20 URLs: http://webapp.example.com/, http://webapp.example.com/login.cgi and http://webapp.example.com/app.cgi?arg=1 through 18. Then I go to http://webapp.example.seanh.net/ and do a few things. Next time I type "webapp" in the location bar, the history will pop up with something like: http://webapp.example.com/ http://webapp.example.com/login.cgi http://webapp.example.com/app.cgi?arg=1 http://webapp.example.com/app.cgi?arg=2 http://webapp.example.com/app.cgi?arg=3 http://webapp.example.com/app.cgi?arg=4 http://webapp.example.com/app.cgi?arg=5 http://webapp.example.com/app.cgi?arg=6 http://webapp.example.com/app.cgi?arg=7 http://webapp.example.com/app.cgi?arg=8 http://webapp.example.com/app.cgi?arg=9 http://webapp.example.com/app.cgi?arg=10 http://webapp.example.com/app.cgi?arg=11 http://webapp.example.com/app.cgi?arg=12 http://webapp.example.com/app.cgi?arg=13 http://webapp.example.com/app.cgi?arg=14 http://webapp.example.com/app.cgi?arg=15 http://webapp.example.com/app.cgi?arg=16 http://webapp.example.com/app.cgi?arg=17 http://webapp.example.com/app.cgi?arg=18 http://webapp.example.seanh.net/ http://webapp.example.seanh.net/goo.cgi http://webapp.example.seanh.net/blah.cgi?num=1 http://webapp.example.seanh.net/blah.cgi?num=2 http://webapp.example.seanh.net/blah.cgi?num=3 http://webapp.example.seanh.net/blah.cgi?num=4 http://webapp.example.seanh.net/blah.cgi?num=5 http://webapp.example.seanh.net/blah.cgi?num=1&letter=a http://webapp.example.seanh.net/blah.cgi?num=3&letter=z http://webapp.example.seanh.net/blah.cgi?num=7 http://webapp.example.seanh.net/blah.cgi?num=8 If I'm trying to get to the top level of http://webapp.example.seanh.net/, this arrangement forces me to type the whole thing out, scroll through the (potentially extremely long) list to find it or find one of the sub-pages of it and delete the part of the URL I don't want. This makes things unusable very quickly if you visit a lot of sites with a common beginning string. It seems like just giving the root URL would be a lot more useful. Or at least sort things by most recent access so that things I haven't hit for days don't get in the way of the one I go to ten times a day. sean -- Sean Harding |"art may imitate life http://www.dogcow.org/sean/ | but life imitates t.v." Address in header *is* valid | --ani difranco
