Comment by tobwen | Hacker News Reader

Comment by tobwen | Hacker News Reader

I don't like the statement that competition in file systems is a good thing. In the end, it means that the users are reluctant because all file systems are either only half-finished or there is an eternal discussion about which one is best or that the other one is broken by design. In the end, only the file system that is used by default by the major distributions wins anyway (currently ext4). This does not mean that no development should take place, but one could simply continue to improve btrfs instead of adding something new that will then bob around in the kernel forever.

We have the same problem in the industry, e.g. with batteries. Instead of the best scientists in the world getting together in a think tank, tanks are used to fight each other. After all, we are all pursuing the same goal, but this is getting bogged down because of profit maximisation, reputation and patents. It's not for the "human race", it's "human against human in a race".

The older I get, the more I wonder whether we need ANOTHER standard or ANOTHER new programming language or ANOTHER new Linux distribution... Maybe just sit down together and work together to improve it.

> this is getting bogged down because of profit maximisation, reputation and patents

This gets worse when there's less competition, not better.

So do you think it makes sense for six to eight different automotive companies to conduct independent research instead of combining their research? In the end, you may have 2-3 good models, which are then either financed with high licence fees or patent penalties.

Sure, and that's why "teamwork" is being forced on students at educational institutions (at least in some states of Europe) :)

Yes, competition generally leads to more innovation and lower prices in industry. Do you know of any economist who thinks otherwise, except in specific cases where natural monopolies exist?

If there's no competition there's no incentive to innovate.

I often argued with business professors during my studies because the theories look good on paper but don't work in reality. If one company invents a new method and patents it, another company will not be able to improve on this method, but would rather try to get round the patent/license costs than really innovate. For me, that is not progress.

But if Linus thinks that bcachefs is stable enough for the kernel, then there must be something to it.

Different filesystem for different things. Sun should have stuck to ufs we didn't need zfs?

In my opinion, this is the case with the "next gen" filesystems: They are actually not THAT different: CoW, reliabity, RAID, performance, etc. - I'm not comparing FAT with ZFS, I'm comparing btrfs with bcachefs.

Reply on news.ycombinator.com