On Sun, Feb 27, 2011 at 17:43, Jonathan Corbet <corbet-ft(a)lwn.net> wrote:
On Sun, 27 Feb 2011 14:38:23 -0600
"Jason D. Clinton" <me(a)jasonclinton.com> wrote:
> What we're trying to do is come up with sane defaults for plenty of good
> user interface reasons. I mean, if we give someone a Linux laptop and the
> first thing we tell them to do is go change a bunch of settings to make
> usable, then we have failed to produce an OS which is worth giving to
> but enthusiasts (have you changed the default fonts on your Nexus One, by
> the way?).
I'm all for sane defaults. Nobody is arguing against sane defaults. But
I spend all day staring at my screen, and I do not believe that "sane"
means "optimal for my particular use of the system."
Please make sane defaults. But please realize that they will never work
The evidence is quite to the contrary. From the body of successful, widely
deployed competition to GNOME is that, indeed, a good default font seems to
work for just about everyone (modulo a11y). We've done our homework, believe
me; here's a copy of *just* the whiteboard of the work done on this:
And that doesn't include all of the discussion and debate that has occurred
among the designers and implementors. And that's *just* the fonts control
We *are* trying hard to get it right. It's like urban planing: it's really
hard to figure out which buildings are the right ones to knock down so that
progress can be made. This just happened to be one of them and one which you
cared for. However, it was likely the right decision.
That isn't to say that the default which has been selected is correct. It
seems correct to me but perhaps you are seeing some kind of DPI-related
scaling issue which is making 11pt quite incorrect. (You should be at 96 DPI
and some xdpyinfo output might be helpful here.)
If you take away the ability to change fonts, I predict
that you will lose a lot of users. That's not sane, that's
better," and it will upset people.
Yes, some people will leave; particularly, power users who are comfortable
spending hours getting their desktop just the way they want it--KDE, xmonad,
and Awesome are all great for that kind of user. But I hope they'll enjoy
the improvement in user experience and that will convince them to stay.
Developers are users, too.
[Along those lines, am I ever going to regain the ability to put the
control key back where $DEITY meant it to be without having to use
I don't know the answer to this question; perhaps Owen can jump in here.
[And yes, I've changed most of the settings on my N1. Several
various reinstalls. The N1 lets me do it.]
Well, I have an N1 and have put custom ROM's on it and utterly failed at
changing the defaults to anything that didn't result in epicly breaking the
rendering of applications from Market. I intended that as evidence support
my position. ;-)
Anyway, if you can provide some screen shots that show that the new
> is not sane, that would be helpful.
Despite my advanced age, my visual acuity is pretty good, but my screen
space is always limited. I can show you that the bigger fonts crowd
things out, force the creation of scrollbars where I had none before, and
generally waste space that I can use better. But why should I have to do
that? Why should I have to convince you that my preferences are "sane"?
Only looking for a DPI-related bug, that's all.
Since you're clearly a bit of a more advanced user, you may be
> changing the dconf key that I referenced in the first link though
> set org.gnome.desktop.interface font-name 'Cantarell 10'` and in the
> when dconf-editor stops crashing continuously, you could do the same
> a GUI tool.
Why that's better than the font selector that we all know well is a
mystery to me. This stuff used to Just Work.
There's two really specific cases where having
yet-another-control-panel-applet is not good: discovery of the settings that
users *should* want to change and, in the support side of things, users who
change the font, don't know what they've done, and then have to call to
$linux_savvy_family_member or $corporate_IT_help_desk. We added the a11y
mechanism to handle vision-related needs specific to fonts in a way that was
simultaneously safe without requiring an entire applet.
It seems like a clear conclusion that someone who actually knows how to
change the font of their entire desktop in a way which could potentially
break their desktop should also know how to fire up dconf-editor.
There's also discussion of creating a "power-users settings
> who are disgruntled by some common defaults but
Such a tool would be a good thing. The alternative is that quite a few of
those users are likely to go somewhere else. Some will do so quite
loudly. You've seen what happens when a desktop environment upsets its
users with a major update; are you really not concerned about repeating
Sorry if I sound strident. I like GNOME, I've been using it for a long
time. I'd hate to see things go wrong in such an unnecessary way.
But I think that most will stay; the Shell is a huge usability improvement
and the way we went about moving from GTK+ 2.x to 3.x has preserved a code
base continuity that KDE could not do in the transition from Qt 3.x to 4.x.
Almost everything, application-wise, is exactly identical to what people are
Are we taking some risk? Sure. Do you remember "10x10?" GNOME has existed
for nearly 15 years and we've gotten, essentially, no where. We have no more
desktop/laptop market share than we had 10 years ago. Looking back at the
origination of the 10x10 goal in 2005, it's really awe-inspiring to think
that we used to be so optimistic while simultaneously being so naive.
We have to, as it is popular in the design community to say, make computer
not suck and so far we've utterly failed at that. I think that most in the
GNOME project have come around to a pragmatic approach in our Linux advocacy
(remember, we're on the same side here!): we're tired of spinning our wheels
repeating the same mistakes over and over again without any measurable
improvement in user experience or adoption. For example, this email has been
passed around inside the GNOME community again and again; it seems to
resonate with a lot of us:
We think that we are making the right decisions and that, hopefully, you'll
finally be able to put a GNOME computer in front of a normal person and not
have them run away kicking and screaming. Maybe 3.0 won't quite be there yet
but we're on the right path, anyway.
I hope this email changes your mind about the removable of the Font capplet
but even if it doesn't, please give Shell a chance; I think you'll find that
it's a net improvement.