[Arny Krueger]

"anahata" wrote in message
et
On Tue, 11 Nov 2008 20:26:24 -0500, Soundhaspriority
wrote:

FAT32 is a fragile file system, not suitable for
archiving.

It's perfectly OK for that, as long as you don't hit the
file size limit. What it's not good for is heavy use with
lots of concurrent file system activity as it's prone to
fragmentation and not particular easy to recover after a
crash or power failure.

FAT32 also has the advantage of wide support as its
specification is freely available.


Furthermore, FAT32 as implemented by most modern USB flash drives has become
something that it was not.

Take a modern 4 GB USB flash drive and plug it into a USB port on a
computer. The computer auto-accesses it, and you can copy files on and off
of the flash drive freely. Nothing special about that, FAT32 has supported
this for years.

Now, even while copying a file onto the USB flash drive, yank it out of its
USB port. Odds are pretty good that at worst, you'll get a warning message
or two.

What's wrong with this picture? Well, if that USB flash drive was a textbook
implementation of FAT32, the drive would now be corrupted. It would probably
be recoverable, but a lengthy scan and fixup would be in order. But, here in
2008, the odds are very good, the drive won't be corrupted. Next time you
plug it in, it will at most be missing the last file you copied onto it.

How can that be?

It turns out that modern flash drives include an intelligent controller that
wraps FAT32 with a journaling file system. So, when you pull it out of the
port and before it is ready the next time you plug it in, it backs out the
last file that was partially copied and leaves the disk whole and ready for
your next usage. This either happens using stored power, or is completed the
next time the drive is plugged in, but before it is made available to you.