Please don’t theme our apps

Web Fonts are a bit different - it's still semantically just text, and you have the option to override with a user stylesheet.

For many websites, web fonts have been an improvement, especially considering the sorry state of "web-safe" fonts back in the day.

I think you're onto something, but I don't think it's as simple as just having a "light theme" and a "dark theme".

I think that if we're going to have toolkits that support custom, nested widgets, those widgets need to be designed to support the context they are in. There will always arise situations where normally-light widgets need to appear on a dark background, or vice versa. The nested path bar example is exactly this - it's a widget that supports a single context (light background), but is having two contexts forced on it by leaky style cascades (light text, but also light background).

How about we bake the concept of "light" and "dark" into the widgets themselves and take advantage of the cascading nature to switch the theme of a widget depending on the context it is in.

Realistically though, GNOME 3 was rather short-sighted when it came to themes. Windows 95 showed us the way 24 years ago: you can change the colours of anything you want. You can tweak the size of most things. You can change the fonts. You can change the icons. Simple API that gets applied on top of an existing theme. Expand this for 2019 with dark and light variants for each theme, and I bet we'd satisfy almost all users.

I have a lot of sympathy for the signers of this letter. As Havoc Pennington (team lead of the GNOME 2 project) liked to point out, every option you provide to users has a cost to the developers in terms of testing and maintenance. Maybe a better way for the letter signers to communicate their position would be to simply say "We don't support nonstandard GNOME themes". Certainly, they're within their rights to say this, and you're within your rights to not use their apps if you don't like it.

Another thing Havoc Pennington liked to talk about in developing GNOME 2 was 'crack'. These were things in GNOME 1 that were bad for the community, but that a lot of users were addicted to. Like, for instance, an overabundance of user settings.

By all accounts, it's hard to write a GUI app for the Linux desktop, in part because of the huge variety of distros and desktops, and apparently even themes within a single desktop. And as a user, my biggest complaint about desktop Linux is the relative paucity of high-quality applications. (Yes, I know there are many high-quality desktop Linux applications. But there aren't as many as on Windows or MacOS.) I can't help but wonder if most of the community wouldn't be better off if a single distro, hopefully with a single desktop and a single theme, decisively 'won' and claimed a dominant usage share of the Linux desktop. Then third-party developers would be incentivized to just develop for that distro, which might make the lives of both users and developers better in the long run.

And of course, since it's open source, if 'the one true distro' went off the rails eventually, someone else could fork it and become the new 'one true distro', with relatively little danger of vendor lock-in.