GNOME Bugzilla – Bug 161872
Use temporal back menu
Last modified: 2017-01-05 22:12:33 UTC
Based on the recommendations of Cockburn et al[1] I believe Epiphany should implement a temporal-based Back-menu instead of a stack-based one. There is one catch, though: the paper suggests that efficiency is higher with a temporal back menu when the menu is used, but actually _lower_ when the user _doesn't_ use the menu but clicks the Back button repeatedly. We could consider the possibility of a stack-based Back function, but a temporal-based Back-menu. Another advantage of a temporal Back-menu is that we can get rid of the Forward button completely, as it (in my opinion) serves no purpose any more. One more consideration: we could also give the History button a temporal dropdown menu, and the Back button a stack-based one. [1] http://www.stanford.edu/group/siqss/itandsociety/v01i03/v01i03a09.pdf
May I suggest http://www.cosc.canterbury.ac.nz/andrew.cockburn/papers/backEval.pdf that is much sorter, and have a diagram on page 3 and a table on page 4 that explains it all. It is a subject that Google have tons of documents about! http://www.google.dk/search?num=20&q=Temporal+back+button
I've been looking up some more information, it turns out that slashdot once had a thread on it: http://slashdot.org/article.pl?sid=02/12/30/1727206&tid=95 A related mozilla bug is https://bugzilla.mozilla.org/show_bug.cgi?id=21521 and also https://bugzilla.mozilla.org/show_bug.cgi?id=187187 Reconsidering the literature, there are still scenarios where there are recently visited sites that can't be reached without the Forward button, so scrap my remark about its lack of purpose for now.
Firefox 3+ uses a combined history browser for Back/Forward and it looks like an ideal use-case for a temporal menu. Therefore I suggest marking bug #525737 as a duplicate of this almost 5-year-old bug.
I have read the paper mentioned in comment #1 and, even if I agree it's an interesting model, IMHO I don't see it better than the current implementation. At least in my case, I like the current stack-based behaviour as for me it's quite natural to blow away the forward history as soon as a new link is selected. For me that means just "new history ahead", therefore new forward history, so current model makes sense for me. Said that (which is just my humble opinion, which doesn't mean I'm against making this change :-)), I guess that we should file a bug in WebKit for this if we want to see it working in Epiphany at some point, as every embed in epiphany relies in WebKitWebBackForwardList objects for the navigation through the Back and Forward buttons.
I'd prefer we not change this. Users will just complain that we're different from other browsers. Stack is familiar and natural....
Wow, has it been 12 years already! At this point I respect this decision and I don't really care any more. However I would like to point out that "users would complain we're different" has never been a good argument in Epiphany history. If they want a different browser, they're free to choose! I believe design decisions should be based on their scientific merits. We're not only catering to users who have been online since 1994 but also tot ones who have never used a computer before.