summarize and hopefully come to conclusions:
1) favicon.ico: not too much *good* discussion. AFAICT there
is no 'official' (as per the W3) standard of indicating an icon file.
I did some research on this, read on.
Apparently recent builds of Mozilla support Website/Shortcut icons using the
<link REL= "icon" HREF= "....."> method. However, (AFAICT) the
W3's HTML 4.01 spec does not indicate any support for an icon.
See the 4.01 Link spec for more info.
However, this MicroSoft Developer Network site
(ya ya I know, Microsoft Sucks...) indicates that the icon
can be designated (*to IE* at least) with HTML.
Since the W3 doesn't appear to support the 'embedded' icon I would
hope that Mozilla follows this guideline as per Mozilla's 'prime
directive' to be, "The most standards compliant browser available."
In any case, the only sensible thing to do is GET the icon if *and only if*
the page indicates and icon exists. ASSUMING the favicon.ico exists
is silly and is a poorly implemented idea.
2) NO-CACHE: not too much good discussion here either. Some say, "turn
it on" others say "wait a bit" and fix the *bigger* problems.
I believe the chaching issue is important to the end user experience,
but to make the caching work properly, other issues must be
addressed. *BE PATIENT* all will be revealed (and working) in time.
:)
3) BACKSPACE Keyboard shortcut: Uggh, so much has been discussed it is
hard to decide what is the best 'default' behavior. It is fairly
apparent that people do not like the BACKSPACE/SHIFT+BACKSPACE
navigation shortcuts. Alternate shortcut combinations have been
offered up, but nothing has been an outstanding choice.
One thing that is fairly unanimous is that the
BACKSPACE/SHIFT+BACKSPACE shortcut keys have to go. It appears
to be too ambiguous (and I agree). Backspace should do nothing but delete
text.
IMHO, I do not see what the problem with (left/right) ALT+ARROWKEY
is, but it was suggested that the coding for this shortcut be placed
in region specific code so that it may be changed (easier) to adapt
to non-US keyboard layouts. Also, I believe ALT+ARROWKEY is probably
the 'least conflicting' key combo on most systems.
I hope someone finds this useful and maybe we can put this thread to rest.
Regards,
J M
Jesse Ruderman wrote:
Netscape recently checked in three changes that had been discussed and shot
down in older bugs. I don't believe that they notified the participants in
any of the older bugs. Each change was checked in within two hours of the
filing of the new bug, leaving no time for users to object and for QA to
mark the bugs as wontfix. If Mozilla 0.9.7 is going to be better than
Mozilla 0.9.6, these changes should be reversed.
FAVICON.ICO
Bug 109843: 14 minutes from filing to checkin.
Mozilla now automatically attempts to retrieve "favicon.ico" the first time
a user visits a site, effectively making it impossible for sites to opt out
of the site-icon feature. If sites do not add a favicon.ico file, they will
still have to pay for extra bandwidth used, and will have to waste time
looking through error logs full of "favicon.ico not found". This is about
as antisocial a browser can be short of implementing bug 109750.
The original summary of the bug was "Turn on favicon pref", which I changed
to "Turn on aggressive favicon.ico search by default" after the checkin was
made. I thought "favicon" (also mentioned in the Mozilla 0.9.6 release
notes) meant supporting site icons in bookmarks, not asking every server for
a file called "favicon.ico".
I did not see an announcement in any of the newsgroups I read. Comment #20
indicates that it was discussed as a subthread of a thread that was 10 days
old at the time (new subthreads of old threads are nearly invisible to most
news clients).
Previous discussion in bug 106328 and bug 32087 showed no indication of a
desire to honor site icons except those linked to with a <link rel="shortcut
icon">.
I have seen numerous complaints about this feature, both in Bugzilla and on
the Mozilla 0.9.6 Slashdot story (which is amazing considering that the pref
was turned on only on the 0.9.6+ tru n k).
NO-CACHE
Bug 101832: 97 minutes from filing to checkin.
Mozilla now reloads pages with no-cache headers when the user hits the back
button. This makes Slashdot threads virtually impossible to read without
opening a new window for each link, even on a fast connection (bug 105395,
wontfix).
Counter-bug 112564 has already been filed.
BACKSPACE
Bug 108816: 31 minutes from filing to checkin.
Mozilla now has the Backspace key mapped to the back button, and
Shift+backspace mapped to the forward button. The shortcut had been removed
more than three months earlier in bug 69981 for at least four reasons:
- It's generally a good idea to avoid having shortcuts that work in one
place but don't work in another place.
- Backspace can't go back a page when you're in a form field, and most
search engines automatically focus a form field when you load the page.
Users who only know the backspace sh o rtcut will be stuck.
- It's easy to "miss" a form field while trying to focus it, and going back
can cause dataloss if the user doesn't realize that the correct way to
return to the page with the form is to use the "forward" button (rather than
clicking the link a second time). Mpt was helping someone who almost lost
an e-mail message this way, and stopped the user just in time.
- Pressing backspace while using a form in a Flash application would cause
Mozilla to leave the page containing the Flash application.
Bug 69981 was reopened.
