Hacker News new | ask | show | jobs
by CD1212 5068 days ago
In my opinion Gnome needs a completely new start, from the ground up.

1. When I last used GTK (about 2 years ago) it felt too big, old and bloated. If GTK were simplified and followed Qt's lead into scripting and easier interfaces (eg. Qt Quick), plus a MIT or LGPL license, this would encourage a new culture of apps.

2. I hated Gnome 3 and Unity for that matter. Gnome 4 needs to take a step back and get out of the way. You don't use the computer just for Gnome, but you use Gnome as a stepping stone. All common apps should be one click away and everything should be as customizable and flexible as possible.

As kljin said, some Gnome apps are redundant and the workforce could do a much better job focusing on the core issues, that could bring more people back to Gnome and hence possibly continue these projects again in the future.

3 comments

> 2. I hated Gnome 3 and Unity for that matter. Gnome 4 needs to take a step back and get out of the way.

> You don't use the computer just for Gnome, but you use Gnome as a stepping stone.

> All common apps should be one click away and everything should be as customizable and flexible as possible.

This is exactly what they chose not to do.

https://mail.gnome.org/archives/usability/2005-December/msg0...

In a reply to Linus Torvalds:

> Sorry to snip mid-sentence, but this is an important point: We're not aiming

> for "powerfully extensible". We're aiming for "Just Works". Some people will

> hate that. Some will love it. Personally, I'd rather have passionate users,

> lovers and haters, than be than average and ignored, and I think you'll find

> most GNOME developers feel the same way.

So they made their choice, ignoring most of their own base were against it. Guess what? they lost users left and right.

They had a sweet spot in Gnome 2 and a strong user base. They thought they could somehow get Windows and Mac users, but that wasn't just happening. At least not at that particular moment in history.

So that was their decision and they didn't backtrack despite all the warnings (plenty during the last 7 years). Stubbornness gets you to obsolescence.

I don't want to bash Gnome. I have Debian testing (Wheezy), and I did try Gnome 3. It's steadily improving, though I still have issues with it. I try it every now and then - but I'm retreating to XFCE for the time being.

Gnome feels un-unified. I guess this has always been the way under Linux, what with QT, GTK and other toolkits, but when one app is slightly at odds with the rest of the Desktop - it feels, odd.

Example being non GTK3 apps, like LibreOffice (though someone here suggests that that is being rectified.) Even VLC, Opera, Firefox and Chrome feel a little out of place. Each behaves differently. You can't quit VLC with CTRL+W for example. Each are designed on different toolkits. Menus are inconsistant. Tabbing behaves differently in each app (can't we relegate this to the Window Manager or Desktop?) Keyboard shortcut unification doesn't exist. These are the edges I'd like to see addressed across the Linux desktop as a user.

Perhaps unification is a lofty target. And we should just be happy with the fragmented cottage patchwork.

I don't even know the difference between GTK, GTK2 and GTK3 and QT! My desktop is such a pain to theme it's a nightmare, I certainly notice that. What's new in layman's terms in GTK3?

As for some core apps, Evolution looked promising. But even that feels a little rough around the edges (I can crash it quite easily.) Thunderbird doesn't integrate with Gnome brilliantly. One flagship email client would be nice.

Focus on the core, the desktop design guidelines and some intrinisically needed apps. Most desktops on Linux seem to suffer in much the same way. Unity still appears ad-hoc.

I guess a good aspiration would be to make it as simple as possible for people to create applications as well as use them under Gnome.

Could there be some kind of CSSification of an app's controls? Present them as interfaces that could be styled differently according to the platform you are on. Leaving Window Managers to take on the role of innovative desktops. Perhaps apps are designed like this already? At least it would be easier to port an application across different form factors.

> "Could there be some kind of CSSification of an app's controls?"

That's exactly a feature that GTK 3 introduced: CSS theming.

All great, except a large portion of apps are still Gtk2 based, which means any potential theme developer needs to make his theme at least somewhat compatible with Gtk2 so there's a consistent desktop.

What should have been done, is the CSS in Gtk3 should've been designed with backward compatibility in mind, and the tooling to automate the creation of a Gtk2 theme based on a Gtk3 stylesheet.

As it is, they can't even move from 3.0 to 3.2 without breaking themes. I've no idea what the situation was for 3.4 or later, but I imagine much of the same.

Also, theming is missing a usable distribution and installation model. Currently, users are expected to simply extract an archive into a specific directory and follow any instructions that come with the theme - some include shell scripts. This is hardly "user friendly," like they claim they're attempting to make Gnome.

How abstract is this though? Can I say take the main menu and do what I like with it? Can I grab the local task menu? Etc.

Rather can this be left completely to the Window Manager? I.e this is a menu, this is a tab. Now present it as you will.

> a MIT or LGPL license

GTK is LGPL licensed.