Re: Question to the candidates.

[Josh Triplett <josh joshtriplett org>]

On Mon, May 25, 2015 at 10:52:55PM -0400, Erick Pérez Castellanos wrote:
> First, thanks to all of you for running as directors.
>
> Currently, GNOME is a strong platform for development, but it's lacking
> integration and features to be a complete, fully integrated desktop
> environment like Mac OS X, for instance. My question is:
>
> "What plans do you have to make GNOME a more complete, fully working
> solution as desktop environment."

First of all, I would suggest that the board of directors has little to
do with the overall direction of GNOME's development, from either a
technical or design point of view.  The many GNOME contributors drive
that.  The board is responsible for ensuring the continued availability
of the resources those contributors need, and for helping maintain and
grow the community, among various other organizational issues, but the
board does not set development directions in that way, nor do I think it
should.

That said, I'd still be happy to answer the spirit of your question.

I think GNOME actually has a huge amount of integration, polish, and
cohesiveness.  And there are several aspects of other environments that
I hope to never see on GNOME; personally, for instance, I don't really
want to see desktop environments delving into "app store"-style package
management.  I like distribution package management just fine.  (I *do*
find sandboxing mechanisms highly appealing for security; I'd just like
to continue installing applications, sandboxed or otherwise, through
apt.)

There are three areas I do think GNOME could use some additional
integration and polish in.

First, while I'm extremely impressed by the huge variety of UI design
ideas that GNOME has experimented with, and many of them have been quite
successful, I think GNOME needs some mechanism to recognize when an idea
isn't working or doesn't really appeal to the majority of users, and say
"well, that was a fun experiment, but let's drop it".  For instance, I
rather strongly suspect many GNOME 3 users would sigh with relief if
alt-tab went back to "switch windows" and alt-` became a
backward-compatibility synonym.  As far as I can tell, though, design
ideas only really tend to get dropped when they get replaced with some
other, newer design idea; for instance, the notification tray gave way
to the excellent new notification mechanism in the latest release of
GNOME.

I think it would make sense to have a convenient way to float design
experiments as extensions or branches, rather than as part of mainline
GNOME, until they become less experimental.  And in the meantime, I
think we need a way, as a community, to decide that a UI experiment was
unsuccessful and should be reverted.

Second, GNOME still needs to improve its support for high-DPI displays.
Right now, GNOME has great support for high-DPI displays that are
sufficiently high-resolution that scaling everything 2:1 is appropriate;
for instance, on a 3840x2160 display, doubling everything and
effectively treating it as a 1920x1080 display works quite well.
However, *many* laptop and desktop displays still have resolutions for
which 1:1 is far too small, but 2:1 is too large, defeating the purpose
of purchasing a high-resolution display.  Treating a 2560x1440 display
as 1280x720 makes things awkwardly large.  Treating a 3200x1800 display
as 1600x900 is still not quite great.  Long-term, ever-increasing
resolutions will likely make the integer-scaling approach viable;
however, in the meantime, supporting only integer scaling leaves much to
be desired.  I've heard comments from several different people saying
that GNOME seems almost cartoonishly big on their displays, and that if
they turn off scaling it's too small.  On top of that, GNOME only
supports a single scaling factor for all monitors, which doesn't work
well when switching between a high-DPI and low-DPI display, or using
both and moving windows between both.

On that front, I actually am taking a concrete step to help there: I'm
donating a high-DPI (2560x1440) laptop to GNOME, to be sent to an
appropriate developer or developers.

And finally, touching directly on your comment about integration, I
don't think GNOME can focus exclusively on its own native applications,
without also taking into account that many users will run non-GNOME
applications side by side with GNOME applications.  As GNOME continues
to produce innovations in UI, desktop integration standards,
system-level features, and similar, someone needs to take the time to
integrate such enhancements into popular third-party applications.  For
example, there has been a patch for Firefox to work with the GNOME
application menu for a year or so, but nobody has stepped up to take
responsibility for that code once merged.  Empathy has deep integration
with GNOME, but Pidgin does not.  There's no default desktop integration
with third-party music player software; that requires a gnome-shell
extension.

These aren't necessarily innovative ideas.  They're not thoughts on how
to make GNOME the desktop of the future, with perfectly integrated
functionality, a suite of native applications, and a novel new UI.
These are points that would help GNOME do better at what it does today,
and take better care of the users it has.

- Josh Triplett