File Properites: Size on disk

[David]

When I selected a folder with files and sub-directories
in it and right-click to get properties, I noticed that
the File Size was like 650mb but the Size on disk was
750mb. What is the difference between the two for Windows
XP. Just so you know I was burning the cd in Nero and it
was 650mb not 750mb.

[Mark Weinreb]

"David" > wrote in message
...
> When I selected a folder with files and sub-directories
> in it and right-click to get properties, I noticed that
> the File Size was like 650mb but the Size on disk was
> 750mb. What is the difference between the two for Windows
> XP. Just so you know I was burning the cd in Nero and it
> was 650mb not 750mb.
>
>

It's all down to the cluster size. Also known as the allocation size. The
actual size of a file on disk is always a multiple of the cluster size. So
if your disk is formatted with 4K clusters, then a one byte file will
actually be 4K in size. The cluster size is always a multiple number of
sectors and can be from 512 bytes (single sector) to 64K on a very large
Fat32 formatted drive.

The difference between a files actual size, and its size on disk is often
called the slack space.

The difference of 100 MB in your case, means that you've got a lot of small
files.

Unless I'm mistaken, CDs use small clusters - they may even use variable
size sectors.

Hope this makes some sense.

Mark Weinreb wrote:

> The cluster size is always a multiple number of
> sectors and can be from 512 bytes (single sector) to 64K on a very large
> Fat32 formatted drive.

Not true (though the first part of your post is correct), sector sizes
are always 512 bytes, formed into cluster sizes that are default under a
FAT** format, going from smaller to larger cluster sizes, according to
different drive/partition sizes!

[Mark Weinreb]

wrote in message news:e5f0d46a71be913dc0946774a55ce277@TeraNews...
>
>
> Mark Weinreb wrote:
>
> > The cluster size is always a multiple number of
> > sectors and can be from 512 bytes (single sector) to 64K on a very large
> > Fat32 formatted drive.
>
> Not true (though the first part of your post is correct), sector sizes
> are always 512 bytes, formed into cluster sizes that are default under a
> FAT** format, going from smaller to larger cluster sizes, according to
> different drive/partition sizes!
>
>

Ahh, but I didn't want to confuse any more than I had to. If you really want
to get picky, then on an NTFS formatted partition, really small files (a few
bytes, say) don't take up *any* space on disk, as they are entirely
contained within the files' directory entry in the MFT.

Mark Weinreb wrote:
> wrote in message news:e5f0d46a71be913dc0946774a55ce277@TeraNews...
>
>>
>>Mark Weinreb wrote:
>>
>>
>>>The cluster size is always a multiple number of
>>>sectors and can be from 512 bytes (single sector) to 64K on a very large
>>>Fat32 formatted drive.
>>
>>Not true (though the first part of your post is correct), sector sizes
>>are always 512 bytes, formed into cluster sizes that are default under a
>>FAT** format, going from smaller to larger cluster sizes, according to
>>different drive/partition sizes!
>>
>>
>
>
> Ahh, but I didn't want to confuse any more than I had to.

Oh dear (LOL)! You were already confused

> If you really want
> to get picky, then on an NTFS formatted partition, really small files (a few
> bytes, say) don't take up *any* space on disk, as they are entirely
> contained within the files' directory entry in the MFT.
>
>

LMAO! no matter what the allocation unit size, a file of a smaller size
will always create slack, it is just less with the 4K default cluster
size used under NTFS. You will get the same results (though less) if you
do the same thing on a folder as you stated in your first reply. Right
click on a desktop shortcut, for example, and you'll see what I mean.
But if you post back, saying that I am wrong, then I will just let you
live in your continued confusion of how sector/cluster sizes don't work.

[David Candy]

It would still use !K for the MFT entry and 16 bytes (or whatever) for =
it's directory entry

I thought if we are going to be precise here then ...
--=20
http://www.smh.com.au/news/opinion/webdiary/index.html
---------------------------------------------------------------
David Candy
http://www.mvps.org/serenitymacros
---------------------------------------------------------------
"Mark Weinreb" > wrote in message =
...
>=20
> wrote in message news:e5f0d46a71be913dc0946774a55ce277@TeraNews...
> >
> >
> > Mark Weinreb wrote:
> >
> > > The cluster size is always a multiple number of
> > > sectors and can be from 512 bytes (single sector) to 64K on a very =
large
> > > Fat32 formatted drive.
> >
> > Not true (though the first part of your post is correct), sector =
sizes
> > are always 512 bytes, formed into cluster sizes that are default =
under a
> > FAT** format, going from smaller to larger cluster sizes, according =
to
> > different drive/partition sizes!
> >
> >
>=20
> Ahh, but I didn't want to confuse any more than I had to. If you =
really want
> to get picky, then on an NTFS formatted partition, really small files =
(a few
> bytes, say) don't take up *any* space on disk, as they are entirely
> contained within the files' directory entry in the MFT.
>=20
>=20

[Ken Blake]

In , Mark Weinreb wrote:

> The difference between a files actual size, and its size on
disk is
> often called the slack space.
>
> The difference of 100 MB in your case, means that you've got a
lot of
> small files.


Actually, no, it has very little to do with the number of small
files. It has to do wit the total number of files of all sizes.
Small files hardly waste more space than larger ones. Regardless
of its size, any file whose size is not an exact multiple of the
cluster size, wastes part of the last cluster it uses--sometimes
a little of that last cluster, sometimes a lot of it. So very
roughly speaking, every file you have wastes an average of half
of cluster.

--
Ken Blake
Please reply to the newsgroup