awhitton@bcara132.bnr.ca (Alan Whitton) (06/15/90)
In article <1990Jun14.165124.15493@terminator.cc.umich.edu>, rees@dabo.ifs.umich.edu (Jim Rees) writes: > You can't have both new features and compatibility. For example, the > //netroot that you like so much breaks plenty of programs. Even the > simplest things, like widening the mtime in the stat struct to 64 bits so > you can have microseconds, will break old programs. The sad thing is that > in most cases, the program was broken to begin with, but its defects only > show up when you try to add a new feature to your operating system. People > don't care about this, though. The only thing they care about is "your > operating system broke my program." Ah but if you look at AFS it looks remarkably like the Apollo File System (an AFS of a different kind). In AFS you can access attached nodes via: ls /afs/foo.edu/bsd4.3 whereas in Apollo you would do ls //foo/bsd4.3 Looks similar? I think so..... Cheers, Alan -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- BNR Ottawa Disclaimer: "This is only my opinion" BITNET: awhitton@bnr.ca OR UUCP: ...uunet!bnrgate!forum!awhitton
rees@dabo.ifs.umich.edu (Jim Rees) (06/15/90)
In article <1990Jun14.190554.1416@bnrgate.bnr.ca>, awhitton@bcara132.bnr.ca (Alan Whitton) writes: > Ah but if you look at AFS it looks remarkably like the Apollo File System > (an AFS of a different kind). In AFS you can access attached nodes via: > > ls /afs/foo.edu/bsd4.3 whereas in Apollo you would do > ls //foo/bsd4.3 But that's just syntax. You still have to have a list of all your nodes ("cells" in AFS terminology) somewhere, although you don't have to mount each one individually. The thing I like about the Domain File System is that it's completely democratic. Each node is both a client and a server, and they all get the same view of the file system. When you plug in a new node, it can immediately see everyone else's files, and everyone else can see its files.