On Friday, November 1, 2002, at 11:14 AM, Rich Morin wrote:
As far as the end user is concerned -- nothing.I have also heard this rumor, but I'm unclear how much of the change would be visible to the user (or even the programmer :-). Aside from being less vulnerable to corruption (e.g., from power failures), what visible effects would you predict?
The comment has been made that a journaled file system is slower... yes and no, it depends upon the underlying hardware and a bunch of other things. With other *nix I have worked with the journaling has no impact on file system access times.
The primary advantage to a journaled file system is in terms of robustness. UFS is notoriously UN-robust. HFS+ is not a lot better. (Thankfully there is DiskWarrior.)
A multi-gig journaled file system will boot dramatically faster than an fsck'd one!
From an application programmer point of view, nothing really changes. The APIs won't (make that... shouldn't) change.
However from a low level systems programming point of view, the entire underlying structure is changed. Depending upon which version is implemented, there are potentially many significant changes that can be made concurrently which will eventually make themselves visible up into the user level -- REAL ACLs being one of particular interest, especially for C2 type enterprise applications. Which brings us back to what was essentially your original question. However, knowing Apple I suspect that any kind of dramatically visible change will be the last thing to happen. But this is the big grey area until Apple makes their direction known.
T.T.F.N.
William H. Magill
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
