On Wed, 2005-12-28 at 16:29 +0100, Kristoffer Lundén wrote: > On 12/28/05, Adam Hooper <adamh densi com> wrote: > > > Am I correct in the assumption that the use case mainly is accidental > > > navigations? > > > > For me, it's far more often accidental window-closings. There's no > > usable way to "undo" those. (With accidental navigations, the "back" > > button brings back my text anyway.) > > > > No it does not always do that. There's plenty of sites out there (no I > don't keep a list) where a "back" does not get you your text back. Those are buggy websites. Epiphany -- like any browser -- tries its best to cope with buggy websites, but there will always be a point when the website just sucks too much. Rule #1 of dynamic web design: don't break the "back" button. > As a side note, if last *result* pages where cached, it's be > completely possible to reshow a last place visited (= undo) when > starting the browser again. Especially since we aren't redoing any > requests or anything, just redisplaying the same HTML. To reliably implement the behavior you suggest, we would need to cache every HTML page visited, along with all images. (Or to provide "undo" support of 10, we'd need to cache 10 HTML pages.) Besides the general icky feeling that comes from such browser behavior (pages can be several megs big), it would be very difficult to explain to the user when he's visiting a cached page and when he's visiting an up-to-date page. Not to mention the confusion between "back," "undo," "forward," and "redo" -- that alone is a conclusive reason to throw the idea away for Epiphany proper. There is absolutely no universal way to recognize a "result" page -- heck, you haven't even defined it. And even if you do suggest a reasonable heusteric, we're getting back to the original problem that the heusteric isn't always right -- which is the reason this whole thread was started in the first place. -- Adam Hooper <adamh densi com>
Attachment:
signature.asc
Description: This is a digitally signed message part