On Sat, 07 Sep 2013 03:32:11 -0400, Warp wrote:

> Francois Labreque <fla### [at] videotronca> wrote:
>> >> Where are they saving money?  The code to fdisk a drive with FAT32
>> >> is still there in the code.  In fact, they had to write even more
>> >> code to check the size of the disk before deciding if they would
>> >> make FAT32 one of the available formats.
>> >
>> > I would guess that they need fewer staff to support fat32, fewer
>> > developers to maintain it, fewer testers to test it, etc.  It all
>> > adds up.
>> >
>> >
>> Why would they need fewer people to support FAT32 by arbitrarily
>> limiting the size of the drives you can use it on to 1/1000th of its
>> full potential?
> 
> Is there a reason why someone would want to use FAT32 instead of eg.
> NTFS?
> 
> FAT32 is significantly slower than NTFS with some operations. (For
> example,
> defragmenting a large FAT32 partition can take over 24 hours, while the
> exact same partition with the exact same data as NTFS takes something
> like 15 minutes to defragment.)

Arguably, NTFS is a journaled filesystem, and journaled filesystems on 
flash media generally burn them out faster.  Wear leveling helps with it, 
but the journal really isn't desirable on that type of media because it 
shortens the media's life.

FAT32 (or vfat) is also desirable for lots of mobile devices (for that 
very reason) and for portability between different operating systems.  
While NTFS-3G provides pretty decent NTFS support in Linux, it's still 
not as mature as the vfat/fat32 support in the kernel (of course, it 
never really will be).

Jim

Post a reply to this message