News for hackers. Stuff that matters.
Re:50 Years in Filesystems: 1984 by pjdesno on Thursday March 06, 2025 @04:05PM (#43285070)
Note that there are some significant non-Unix filesystems along the way. In particular, a few years later: "Reimplementing the Cedar File System Using Logging and Group Commit", Robert Hagmann, Xerox PARC. SIGOPS Operating Systems Review, November 1987.

That's the origin of the journaling approach used in post-FFS unix file systems through the 90s. (I don't count ext2 as "post-FFS", as it's basically FFS on a disk with LBA addressing)

Re:50 Years in Filesystems: 1984 by gjvc on Thursday March 06, 2025 @05:32PM (#43285713)
Xerox PARC strikes again
Re:50 Years in Filesystems: 1984 by pjmlp on Friday March 07, 2025 @03:23AM (#43288230)
Still waiting for the Mesa, Cedar, Smalltalk, Interlisp-D set of approaches software development, OSes and systems programming to take over, but I guess we can't have everything at once, and should be happy with what we already got into mainstream out of Xerox PARC.
Re:50 Years in Filesystems: 1984 by gjvc on Friday March 07, 2025 @06:23AM (#43289277)
They don't need to take over in order to be a wild success. I'm reminded of the 1997 speech from Steve Jobs where he said "We have to let go of this notion that for Apple to win, Microsoft has to lose".

Perhaps AIs will recognise their value and convince people! :-)

Re:50 Years in Filesystems: 1984 by convolvatron on Thursday March 06, 2025 @08:05PM (#43286610)
wow. I always thought this was due to M. Rosenblum in 1992. Looking at the paper now it doesn't reference this one. I wonder what other group ideation happened in the late 80s around this. I guess it was already well established in databases.
Re:50 Years in Filesystems: 1984 by Upvoter33 on Friday March 07, 2025 @06:33AM (#43289335)
Journaling and log-structured file systems are different (but related) things