Ask HN: Why don't we have a shared "libchrome" the way we have glibc or DirectX?
3•omagdy7•1h ago
The status quo of shipping a cross platform desktop app easily for the past years have been electron, but this comes at the cost of pretty much shipping a chromium with each app which is not ideal. So why don't we have a libchrome like we have glibc in Linux or DirectX in windows. Imagine every game would bundle their own DirectX version with them. The big blocker I imagineis stability glibc doesn't change much but chrome is constantly changing, but I feel this is easily fixable by maintaining a LTS version of the library and calling it a day. Has anyone pitched this idea before. Is chromium as an engine too coupled in logic with chrome as a desktop app making it impossible to ship it as a library?
Comments
akerl_•1h ago
> I feel this is easily fixable by maintaining a LTS version of the library and calling it a day.
I think you're really underselling the amount of work involved there. What makes you think that's ~"easy".
omagdy7•1h ago
Well yeah I guess "easy" wasn't the most appropriate word here I just wanted to communicate that API stability shouldn't be like a big wall if we wanted to bundle chrome as a library. but yeah maintaining any thing of the size of chrome would never be easy
akerl_•1h ago
There's not really a great framework for GUI apps on Mac/Windows to generically interact with a shared library. DirectX and friends are sort of an anomaly here.
So you'd need somebody to build the tooling for that kind of resource-sharing and convince people to use it (think about similar prior waves where various toolchains became common practice for things like update-checking, installers, etc). And they'd basically have to be doing it out of love-of-the-game, because there's not really any money in it and any time it breaks people would be pissed.
akerl_•1h ago
I think you're really underselling the amount of work involved there. What makes you think that's ~"easy".
omagdy7•1h ago
akerl_•1h ago
So you'd need somebody to build the tooling for that kind of resource-sharing and convince people to use it (think about similar prior waves where various toolchains became common practice for things like update-checking, installers, etc). And they'd basically have to be doing it out of love-of-the-game, because there's not really any money in it and any time it breaks people would be pissed.