Merging bcachefs

Posted Jun 22, 2023 8:04 UTC (Thu) by nim-nim (subscriber, #34454)
In reply to: Merging bcachefs by khim
Parent article: Merging bcachefs

> Sorry, but that's failure in my book. It puts you into a stark bind: either you avoid most desirable (and most lucrative!) employers or you can use the code you wrote.

Once upon a time listing all the companies that made and sold BSD-based appliances was a who’s who of the industry.

Once upon a time Java was the future and the whole big data stack was built on Open Source (*not* Free Software) Java bits (Hadoop, etc). Now Python (definitely not Free-Software hostile) stole the big data show.

All those ecosystems have dramatically shrunk, the companies that remain mostly release Linux and Free-Software based products.

Open Source (as opposed to Free software) is a dead end that gets repeatedly estinguished by the greed of companies that insist on it. So maybe it pays more short term. Long term it condemns its participants to the anguish of being replaced by something else whose proponents share more with one another.

The corporations themselves do not care overmuch one way or another. Change of tech, change of techies. The only ones fooling themselves OSS vs Free Software matters for business are those techies. In the end money talks (Google’s awowed distaste for the GPL won’t make it dump Linux in GCP or Android as long as it makes them money; Microsoft has seen the *lucrative* side of cancerous GPLed Linux a long time ago).

to post comments

Merging bcachefs

Posted Jun 22, 2023 20:05 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

> Now Python (definitely not Free-Software hostile) stole the big data show.

Uh... Whut? Python is quite explicitly OpenSource, not Free Software. It has a permissive license, and there's nary a GPL-ed package in sight. All the recent AI work is built on permissively licensed packages.

> Open Source (as opposed to Free software) is a dead end that gets repeatedly estinguished by the greed of companies that insist on it. So maybe it pays more short term.

What you're saying is completely false. It's Open Source that wins long-term because it's a superior development model. Proprietary forks (and models like "open core") can exist for a while, but they eventually either become obsolete and die off, or get integrated into the open code.

This is doubly so when the software gets complicated. You can't really develop a meaningful proprietary fork of Kubernetes, simply because your "secret sauce" changes will likely be insignificant compared to the overall functionality. And if they are significant, you'll have to keep up with the firehose of new functionality that is released for each new official version.

And finally, Open Source allows companies to collaborate on non-core functionality. My previous company worked in the solar industry, and we as hell wouldn't release our proprietary simulation algorithms, but we contributed quite a few patches to miscellaneous projects (because why not?).

To paraphrase an old Soviet joke, the goal of FSF is to make sure that there's no proprietary software, the goal of Open Source is to make all software open sourced.