Vista theme editorial and late postmortem

[Alex Faaborg]

Editorial about our Vista theme: "Firefox, you're kind of ugly"
http://www.neowin.net/news/main/09/02/03/bang-on-firefox-youre-kind-of-ugly

main points of the editorial:
-bulky, use of a menu bar
-"cheap knock off of a Vista minded theme"
-"It takes the Vista blue menu color and applies it to all of the
toolbars at the top of the window"

We never really did a postmortem on the Vista theme, so here are some
aspects of our process that I think we could probably improve in the
next release (now that the next release is soon to be upon us):

1) We collectively thought of Vista as another "Windows theme", as
opposed to a completely new platform. Imagine if we had just painted
the surface of our XP theme for OS X, but hadn't made any other
significant structural changes: it similarly would have been
considered a cheap knock off by mac users. (It also ironically
probably would have looked kind of like Safari currently looks like on
Windows).

2) We probably over valued cross platform UI consistency. While we
switch between operating systems regularly, this is unlikely the case
for the majority of our users on Vista. I believe Firefox's ability
to mesh with other applications on Vista is more important that
keeping the Firefox UI consistent with the window layouts on XP or
Linux. Previously we have enforced cross platform consistency because
we believe it is a more adoptable interface for users. I'm not
against adoptability, but long term use is considerably more, well,
long term.

3) We were collectively reluctant to step away from the design we were
automatically extracting from the OS and start hard coding images to
proactively control the design. This significantly limited our
options, and we ended up shipping the visual design that the OS gave
us, instead of the visual design that we created.

4) We allocated resources based on the demographics of our user base,
instead of the demographics of journalists. This wasn't necessarily a
mistake, but is a significant part of the process leading to the
current VIsta theme. Given that the vast majority of our users are on
XP, we focused our efforts on making it better (with the efforts of a
single windows team being allocated between the two, OS X was a
separate team). However, journalists aren't likely to be running the
OS from the turn of the century, leading to them thinking we sort of
ignored Vista, which is arguably accurate.

Anything else that we should include the postmortem? Based on these
points the changes we need to make are:

1) Consider Vista and Windows 7 to be new platforms, instead of new
versions. We can't just visually skin for them, we have to deeply
consider how to mesh with their interactive design.

2) Value local integration over cross platform consistency in some
cases, and in general more. This is still going to be a careful
balancing act, since it's not that case that one is clearly more
important than the other, but we can't completely ignore local
integration.

3) Get more detailed OS theme detection and start to hard code to
specific OS themes the same way we approach development on OS X. We
still need to be able to fall back to classic or third party themes,
but requiring the default theme to fall back, or attempting to be
future proof, should both be non-goals.

4) Begin to transition resources from XP to Vista, since it is now in
need of the most work.

Other suggestions?

Cheers,
-Alex

[Axel Hecht]

I guess I'd be interested to learn what you mean by "new platform". In
terms of development, build, l10n, testing.

Another thought I'd have is that I just started to read a review of
Windows 7. Is it "like Vista" enough to throw them into one bucket? If
not, is Vista worth going for? I know that lots of users didn't.

Axel, who's hardly using windows these days anyway.

[Dao]

I don't think there's something deeply wrong with how we handle Vista
with regards to the theme infrastructure, native theming vs. making up
our own widgets etc.
What I take from the article is that we should use aero glass and that
we need to figure out what to do with the menu bar, but we already knew
that during the Fx3 development.

dao

[Alex Faaborg]

> Another thought I'd have is that I just started to read a review of
> Windows 7. Is it "like Vista" enough to throw them into one bucket?
> If not, is Vista worth going for?

My impression is that for the most part we could group those two into
the same bucket.

> I guess I'd be interested to learn what you mean by "new platform".
> In terms of development, build, l10n, testing.

I was primarily thinking in terms of the mindset in which we frame UI
decisions (which unfortunately due to technical limitations could also
have broader implications). When we thought about UI changes to other
platforms, the answers always seem pretty obvious to us, when we
framed each question in the context of what was right for that
specific platform. For instance:

-Should our OS X version have a menu bar at the top of the window
itself, instead of the top of the screen? ...that would be obviously
wrong
-Should we use native GTK icons on Linux, or just ship a copy of our
XP theme? ...native GTK icons is obviously better
-Should we use a unified toolbar on OS X? ... clearly, every other app
is using a unified toolbar.

So what I'm getting at is that I think we need to mentally approach
Vista/7 as being as different from Windows XP as OS X is different
from Windows XP.

For instance, our main toolbar on Vista is currently a kind of cold
blue-ish-silver color right now. But if you get a fresh install of
Vista, it is very hard to find a similar looking toolbar anywhere on
the system (I'm actually a little curious if anyone can find one).
We are automatically painting the "native toolbar" but we are also
showing the user something that is otherwise completely unusual and
unique. So the question becomes, is that toolbar really native?

If you think of Vista as just another version of Windows, the answer
to "is the toolbar really native" is probably going to be yes, after
all we are drawing it by extracting from msstyles. If you think of
Vista as a new platform, the answer is probably no, the OS doesn't
display toolbars like that anywhere else, regardless of the legacy API.

Now back to the broader technical implications, having a structurally
different UI for Vista (for instance, not using a menu bar) would
require us to detect the OS on startup, which is something that I
believe we currently can not do. So we would obviously either need to
be able to detect the OS, or actually have to ship a different build
for Vista/7 if we wanted to make any modifications to the interface
that weren't just skinning the surface.

Or we could modify our XP interface in order to also fit in on Vista/
7, but personally I think that is as logical as modifying our XP
interface due to a change on OS X.

-Alex

> _______________________________________________
> dev-apps-firefox mailing list
> dev-apps...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-apps-firefox

[Alex Faaborg]

> but we already knew that during the Fx3 development.

yeah, none of this discussion is really all that new, just reviewing
so everyone is all caught up :)

> we should use aero glass and that we need to figure out what to do
> with the menu bar

Can we dynamically change things on startup depending on the OS
version right now? I thought this would require changes to our theme
infrastructure, but it is also entirely possible that I'm confused.

-Alex

[Robert Kaiser]

Alex Faaborg wrote:
> Editorial about our Vista theme: "Firefox, you're kind of ugly"
> http://www.neowin.net/news/main/09/02/03/bang-on-firefox-youre-kind-of-ugly
>
> main points of the editorial:
> -bulky, use of a menu bar
> -"cheap knock off of a Vista minded theme"
> -"It takes the Vista blue menu color and applies it to all of the
> toolbars at the top of the window"

I'm not sure we should take those things too heavily, IE7 looks ugly as
hell and non-integrated on XP, and that seems to not be taken as so much
of a problem. Not to say to we couldn't or shouldn't be better than IE
here, but I wonder how much it is worth the effort.

> 3) We were collectively reluctant to step away from the design we were
> automatically extracting from the OS and start hard coding images to
> proactively control the design. This significantly limited our options,
> and we ended up shipping the visual design that the OS gave us, instead
> of the visual design that we created.

I think that was one of our positive sides on Windows and Linux, and one
thing where we badly suck on Mac. Hardcoding stuff and not going through
stuff like nsITheme always feels strange to me, given that all modern
OSes support more than just a single theme and we would/will/do look
awkward very fast when the user uses a different theme than the
OS-default (of course, the question is how many users are actually using
a different one).

> 4) We allocated resources based on the demographics of our user base,
> instead of the demographics of journalists.

Firefox isn't a browser for journalists but the mass of the users out
there, right? ;-)

> 1) Consider Vista and Windows 7 to be new platforms, instead of new
> versions. We can't just visually skin for them, we have to deeply
> consider how to mesh with their interactive design.

This may be a good idea anyway because of the weirdness of what we have
in winstripe right now with doubled files for all icons and some other
files, but of course more forking makes maintaining the themes and
components using them harder.

Robert Kaiser

[Patipat Susumpow]

Ribbon in Firefox 4?

Sent from my iPhone

[Dao]

On 06.02.2009 10:32, Alex Faaborg wrote:
>> we should use aero glass and that we need to figure out what to do
>> with the menu bar
>
>
> Can we dynamically change things on startup depending on the OS version
> right now? I thought this would require changes to our theme
> infrastructure, but it is also entirely possible that I'm confused.

We already have browser-aero.css, and there we'd add the glass effect,
if it worked for the main window.
We should also be able to detect Vista in Javascript and do the menubar
thing, which is less of a theme question anyway.

dao

[Dao]

On 06.02.2009 10:21, Dao wrote:
> we need to figure out what to do with the menu bar

I've filed bug 477256 for implementing menubar auto-hiding in Toolkit.
Once that's done, we can use it in Firefox.

By the way, generally speaking, I think hiding a menubar until Alt is
pressed feels rather clumsy. But making room for the content area is
really important for us, and if the clumsiness is okay for Vista, we
should at least consider this change for XP as well