Reiser4 Still Lacks Mainline Ambitions
Phoronix: Reiser4 Still Lacks Mainline Ambitions
Reiser4 is still not ready for integration into the Linux 3.3 kernel nor has the file-system even been officially updated yet for any of the recent kernel releases...
no one wants to use murderFS?
No surprise here. Might as well port it to FUSE, it can be another novelty filesystem.
This quote from kernelnewbies sums it up.
"There are not a lot of people who are trusted to do this work. There are even less who are willing to do it."
Trusting just one person is hard. Nothing short of impossible. Really.
The Linux kernel team simply isn't interested in Reiser for the reasons we all know. It's not a matter if it's a good filesystem or not, it's just that there aren't developers willing to enter the "reiser" world.
My 2 cent marketing suggestion for Reiserfs: rename the filesystem to a new name, wait like 2 years so that everyone learns the new name and forgets about "reiser's" history. This is the only chance to get over the political reasons that is blocking Reiser.
A rose by any other name smells just a sweet.
I don't think a name change is going to accomplish much. The issue people have with reiserfs is not a marketing one.
So it looks like it's going to remain a out of tree project. It's not the only FS needing to do this, you know, of course.
The only thing that interests me about Reiser4 are the things that were promised but never happened: accessing files as directories and using this interface as a standard way of managing metadata, putting different permissions on each line/field of /etc/passwd, all the stuff that was in the "future vision" whitepaper.
Without those features what makes Reiser4 more compelling than Btrfs?
The former isn't particularly useful, easier implementation of this would be to extend nautilus, which already has the ability to display file metadata -- just needs to be modified to ALTER that metadata, which is a fairly simple task.
Originally Posted by Abn0rmal
The latter can be accomplished by switching to a database instead of using /etc/passwd. I can't see any particularly compelling reason to build your filesystem around a specific file.