Re: file question
dadiOH wrote:
> > Sometimes the "size on disk" of a file is much larger than the
> > "size" of the file. How can that be avoided?
> Files normally use more bytes than the actual file size because
> the space for the file is allocated in pre-set chunks. How big
> those allocation units are depends on the size of the hard drive.
As dadiOH says, file space is allocated in chunks that are known as
clusters or allocation units. All versions of FAT (and NTFS) allocate
disk space based on clusters that have a fixed size.
When a drive (or more correctly, a FAT-32 volume) is formatted, the
cluster size is set according some built-in rules used by the
format.com program. Those rules are designed such that the cluster
size is adjusted so that you don't end up with more than 2 million
clusters on a given volume.
Microsoft designed it that way so that drive defragmentation and file
allocation table checking wouldn't take a long time nor a lot of
system memory. What that means is that a volume of a given size will
end up with a particular cluster size, as follows:
up to 8 gb -> cluster size 4 kb
up to 16 gb -> cluster size 8 kb
up to 32 gb -> cluster size 16 kb
up to 64 gb -> cluster size 32 kb (max size for a cluster)
For volumes larger than 64 gb, the cluster size remains at 32 kb (the
max size for a cluster) but the cluster count is allowed to exceed 2
million, and reaches 4 million for volumes that are 128 gb in size.
The 4-million-cluster-count number, and the 128 gb volume-size number,
have special meaning or specific explanations (and significance or
consequences) that I (and others) have discussed in previous posts.
The FAT-32 specification does not call for or require increases to the
cluster size as the volume size increases. As a result, it is
possible to format a volume and use a different cluster size than the
format.com program would have chosen.
About 6+ months ago, I posted a series of messages exploring the
consequences of using "non-standard" cluster sizes while running
windows-98. The short answer is that yes, it is possible, and safe,
and certainly more efficient (from a disk utilization standpoint), but
there are some issues that need to be understood (and can be dealt
with).
As to your other question - the difference between "FAT" and "Fat-32",
that has already been answered, but generally I believe that "FAT" is
used as a generic term to refer to the "FAT family" (FAT-12, FAT-16,
and FAT-32) and that it may have been used many years ago to refer to
FAT-12 before the development or deployment of the other 2 FAT types.