Post #66,306
12/3/02 12:00:37 AM
|
Can't be resized?
Yes it can.
If you're running FAT32 in preference to NTFS, then yes, you *do* need FS lectures, sonny.
Why bother? Faster and more robust, that's why.
Peter [link|http://www.debian.org|Shill For Hire] [link|http://www.kuro5hin.org|There is no K5 Cabal] [link|http://guildenstern.dyndns.org|Blog]
|
Post #66,308
12/3/02 12:09:31 AM
|
Re: Can't be resized? - no
Not unless you have an NDA with MS. It's not open. It's the openness, stupid.
-drl
|
Post #66,309
12/3/02 12:10:37 AM
|
Sure it can.
PartitionMagic 7 resizes NTFS partitions.
You can resize partitions yourself if you're using dynamic disks.
But so what? If you wanted *open*, you wouldn't be running *FUCKING WINDOWS*, would you?
Peter [link|http://www.debian.org|Shill For Hire] [link|http://www.kuro5hin.org|There is no K5 Cabal] [link|http://guildenstern.dyndns.org|Blog]
|
Post #66,315
12/3/02 12:29:34 AM
|
PQ has NDA with Borg
I want open. I run Windows and Linux. We all live together in a sonorous harmony.
-drl
|
Post #66,317
12/3/02 12:39:30 AM
|
Re: PQ has NDA with Borg
I want open. I run Windows and Linux. We all live together in a sonorous harmony. That's a bit like saying "I want to sit in the tiger's cage, but I don't want it to eat me." Get over it. Windows is a proprietary operating system. NTFS is its FS. FAT32 is legacy, and it's shite. You chose this path.
Peter [link|http://www.debian.org|Shill For Hire] [link|http://www.kuro5hin.org|There is no K5 Cabal] [link|http://guildenstern.dyndns.org|Blog]
|
Post #66,314
12/3/02 12:27:55 AM
|
Not faster either
In fact, DOS FAT with its enormous clusters and no user context is the fastest FS of all. Speed is not the issue. It's not faster.
Robust? In what sense? Because only NT can r/w it, you need to be running NT to repair it. But how, if the install is dead? FAT32 can be repaired from a boot floppy (as can Linux and ext2).
-drl
|
Post #66,319
12/3/02 12:42:17 AM
|
Faster.
Yes, faster.
Whether you like it or not, NTFS is faster.
Don't believe me? Try copying a 50GB file between NTFS partitions. Now do it between FAT32 partitions.
Peter [link|http://www.debian.org|Shill For Hire] [link|http://www.kuro5hin.org|There is no K5 Cabal] [link|http://guildenstern.dyndns.org|Blog]
|
Post #66,322
12/3/02 12:53:02 AM
|
Re: Faster.
50GB? One file? What kind of benchmark is that?
I can see the little folders flying...
NTFS is slower in the real world, with lots of fixed-place programs and small files, like browser cache and secretary letters.
-drl
|
Post #66,326
12/3/02 12:58:57 AM
|
Whatever.
The real world is people doing digital video of their kids.
The real world is 80G of MP3, Ogg and WMA files.
The real world is where NTFS outperforms FAT32 every fucking time.
Just because you WANT it to be slower doesn't make it so.
FAT32 is dead. Get over it.
Peter [link|http://www.debian.org|Shill For Hire] [link|http://www.kuro5hin.org|There is no K5 Cabal] [link|http://guildenstern.dyndns.org|Blog]
|
Post #66,411
12/3/02 11:31:22 AM
|
Re: Whatever.
Do you think I was born yesterday? I've been dealing with NTFS since you were in school! Simple common sense tells you that NTFS is GUARANTEED slower because of the security and journaling overhead *in most real world situations*, assuming the same level of fragmentation and similar disk layout (cluster size etc). Unless the disk is badly fragmented, FAT32 results in less head motion (you can see this with your own eyes). NTFS has an advantage in speed of directory listings because of the MFT - thus if many, many tiny files are stored in a complex directory structure, with at least some of those directories having hundreds and hundreds of entries, there can be a significant performance increase. But who has a disk setup like that? For an applications-oriented disk, this *single* speed advantage of NTFS is moot. The only case I can think of where this might be significant is say a browser cache that is enormous on a machine with a high-speed network link.
For very large files, the actual file system is almost irrelevant and the main issues are cluster size and fragmentation.
-drl
|