NTFS write support?

[98Guy]

Fragmentation is obviously more probable when the "minimum unit" (the cluster in FAT filesystems) grows in size.

NTFS indexes single sectors, no matter size of volume, thus it is more efficient with bigger volumes, but has other drawbacks.

Note that the behavior of cluster-size increasing along with volume-size is NOT a necessary characteristic of FAT (at least not FAT-32).

There is nothing in the FAT-32 spec that demands that cluster-size increase as volumes get larger. That phenomena is a characteristic of Microsoft format.com utility, which enforces a rule that tries to prevent cluster-counts from exceeding 2 million on any given volume, and will scale up the cluster size to enforce that rule.

I've used third-party drive and partition software to create FAT-32 volumes as large as 500 gb, with 4 kb cluster size (same size as NTFS) and have installed and run Win-98se on such volumes. While it is true that some software (scandskw, defrag, norton utilities) can't handle volumes with "larger than normal" cluster-counts, their limits are generally several times higher than the default 2 million, and win-98 itself has no problems when the cluster count rises from 2 million into the 10, 20, even 40 million number.

[jaclaz]

@98Guy

As said, I am not interested in yet another FAT32 vs. NTFS flamewar, if you are happy with your modified with third party utilities FAT32 and 500 Gb volumes, I am happy for you.

However, NTFS cluster size can be 512 bytes (up to 2 Tb), the 4 kb is the default value for drives bigger that 512 Mb when using format.com without specifying a cluster size.

http://support.microsoft.com/kb/140365/en-us

jaclaz

[98Guy]

As said, I am not interested in yet another FAT32 vs. NTFS flamewar.

However, NTFS cluster size can be 512 bytes (up to 2 Tb), the 4 kb is the default value for drives bigger that 512 Mb when using format.com without specifying a cluster size.

This is not a flame war.

This is just a clarification.

It's a popular myth that FAT-32 *requires* the cluster size to increase on large volumes (resulting in inefficient storage for small files).

I'm pointing out that for win-98, and especially for NT-based OS's like 2K and XP, that they can work with pretty much any cluster-size (even 512 bytes) on FAT-32 volumes (regardless of volume size). That you must use (readily available) tools other than from Micro$oft to prepare and format the volumes in that way is a secondary issue, and has no bearing on the merits or capabilities of the FAT-32 file system.

[dencorso]

I guess my posts #3 and #4 in this topic might interest you: hdd size limits? for ME defragger in MDGx' 98SE2ME.

BTW, I know I can use fat32format in Win XP to acheive almost any partition-size/cluster-size ratio that I may want. I also know that, theoretically, I should be able to do the same using Win ME format with the undocumented /Z switch, but it almost never works for me, because it refuses to do its magic on removable media. What other programs are there that can be used in Win 98SE to do it, instead of having to go to Win XP to format and then go back? Also, do you know of any formatting utility capable of letting me control cluster-size for FAT-16?

[98Guy]

I guess my posts #3 and #4 in this topic might interest you: hdd size limits? for ME defragger in MDGx' 98SE2ME.

BTW, I know I can use fat32format in Win XP to acheive almost any partition-size/cluster-size ratio that I may want. I also know that, theoretically, I should be able to do the same using Win ME format with the undocumented /Z switch, but it almost never works for me, because it refuses to do its magic on removable media. What other programs are there that can be used in Win 98SE to do it, instead of having to go to Win XP to format and then go back? Also, do you know of any formatting utility capable of letting me control cluster-size for FAT-16?

I have just read your link. I think I will repeat some of my tests at some point soon.

The /Z switch is useless - it won't let you specify anything other than the default cluster-size for a given volume size.

As for software, I've used the OnTrack clone software that's been re-branded for use on Seagate and Samsung hard drives. For example:

http://www.seagate.com/www/en-us/support/d...oads/discwizard

Or the Samsung version HUTIL (ver.1.21)

[pinecloud]

Entered by an error - by pressing tab key

Edited December 14, 2008 by pinecloud

[pinecloud]

Replying to NTFS write support?

NTFS Kernel Mode Extension Module Loaders

Winternals (Year 2003)

NTFSWindows98.exe (Install Programme) 1164KB Compressed Archive

Supports NT 3.5 4.0, Win 2000, 2003 and XP

You have to make a subdirectory i.e., "NTFSDrvr" in Windows 98 directory and place the 7 files from your NT

(it tells the name of files when you decompress the installer programme.

Since it uses part of true NT kernel driver, reliability is excellent

It adds Windows 98 Explorer an ability to manage NTFS partitions

just like own FAT32 partition

Paragon (Year 2006)

PNTFS.exe (Install Programme) 2034KB Compressed

Enhanced kernel mode extension driver included that enables

support for Vista, XP, Win 2000, 2003 and NT4.0

FAT32 Kernel Mode Extension for Windows NT 4.0

Winternals (Year 1999)

FAT32.exe (Install Programme) 781KB Compressed

Since it uses part of true FAT32 Kernel mode extension driver from

true Win2000, reliability is excellent

Universal FASTFAT (Year 1998) Prepared for service pack to NT but withdrawn for some reason

FASTFAT.sys 134KB

FS_REC.sys 7.5KB

Replace existing Kernel mode FAT drivers of the same name to gain native access to FAT32

pinecloud,

[SlugFiller]

Unicode File Names

FAT32

System Character Set

Man, who wrote this crap?

FAT traditionally uses UTF-7 for filenames, not a system-variant "System Character Set".

The VFAT expansion adds LFN which are stored in UTF-16, meaning true unicode (well, depending on whether it supports surrogate pairs, though that's implementation specific really).

So, no version of FAT uses "System Character Set".

Most of the rest is misinformation as well...

[jaclaz]

Unicode File Names

FAT32

System Character Set

Man, who wrote this crap?

FAT traditionally uses UTF-7 for filenames, not a system-variant "System Character Set".

The VFAT expansion adds LFN which are stored in UTF-16, meaning true unicode (well, depending on whether it supports surrogate pairs, though that's implementation specific really).

So, no version of FAT uses "System Character Set".

Most of the rest is misinformation as well...

But WHO are you actually quoting?

FYI:

http://msdn.microsoft.com/en-us/library/ms776428(VS.85).aspx

http://msdn.microsoft.com/en-us/library/ms776463(VS.85).aspx

UTF-7 is another thing:

http://en.wikipedia.org/wiki/UTF-7

jaclaz

Edited December 14, 2008 by jaclaz

[erpdude8]

NTFS is crap. There is no practical need for it for the home or SOHO user.

totally out of line, 98Guy (and far from the truth, IMHO). I use NTFS on some of my home PCs and it's very good.

I really do NOT care which file system you guys are using as long as you're happy with it, regardless of whether it's FAT or NTFS.

for DOS sessions, I use Avira NTFS4DOS Personal edition.

[erpdude8]

Paragon (Year 2006)

PNTFS.exe (Install Programme) 2034KB Compressed

Enhanced kernel mode extension driver included that enables

support for Vista, XP, Win 2000, 2003 and NT4.0

Paragon's NTFS for Win9x software (PNTFS.exe) is now free for home users. I've just received a license key from them via a email and the full download, not a demo.

Universal FASTFAT (Year 1998) Prepared for service pack to NT but withdrawn for some reason

FASTFAT.sys 134KB

FS_REC.sys 7.5KB

Replace existing Kernel mode FAT drivers of the same name to gain native access to FAT32

and where exactly can you find these "universal FASTFAT" drivers? Googling them doesn't seem to work and leads nowhere.

[dencorso]

Universal FASTFAT?

Would that refer to the drivers mentioned by bearwindows? Read the "Solution 2" to "Problem 7" here.

Also this thread may have some additional info.

Edited December 26, 2008 by dencorso

[CharlotteTheHarlot]

Paragon NTFS4WIN98

Utter cr@p!\\\\\\\\\\ (politically correct version : use of that product may require a certain amount of caution.)

$0.00 is overpriced, it's not worth the download IMVHO.

In my experience :

- did not identify ANY partitions - not just NTFS ones - on an external USB disk, which however were perfectly mounted and accessible, by letters, in Windows. I gave Paragon NTFS the benefit of doubt and retried twice, wensuring the USB disk was plugged in to the system before rebooting the machine. Still no joy.

- on the main (fixed) disk : failed to create a new NTFS inside of an extended partition, or - actually - it created one but failed to properly chain it to the parent EMBR.

- generally PNTFS conflicts with Vadim Burtyanski's excellent Letter Assigner, a must if you ask me.

- last but not least, while evaluating PNTFS, it crashed Windows 98 SE.

I wouldn't dare use PNTFS to /read/ an existing NTFS part if by chance it succeeded to recognise it, even less so try to /write/ !

Needless to say that thing was removed from my system in less time than it took to install. Good riddance.

If someone /needs/ to access NTFS partitions from Windows 9x, there are other solutions - formerly from Sysinternals' Russinovitch - that actually /work/

Just my 2 cts

--

Ninho

Wow, I guess this qualifies as a bad review! Is this the Paragon free Win9x NTFS read/write driver mentioned in this post and available from here??? Does anyone else have any experience with this on Win9x? I downloaded it and will eventually try it on a Win9x box but reviews such as this tell me to wait a bit. Anyone?

P.S. @Ninho: any chance you could explain the testing? Specifically, I am wondering whether you tried any internal ATA drives or just the external USB? Also, how large were the drives involved? What was the Win9x RAM amount? What was the WinXP SP(?) system used to format the NTFS disk(s) and how was the format/partition done? Did you save the Win9x bootlog.txt? Just wondering. Thanks!

[98Guy]

NTFS is crap. There is no practical need for it for the home or SOHO user.

totally out of line, 98Guy (and far from the truth, IMHO). I use NTFS on some of my home PCs and it's very good.

Totally in line.

The purpose of NTFS was (a) for MS to have a completely proprietary (and largely undocumented) file system with no linkage to FAT-x that they envisioned would eventually be adopted (licensed) for use by other devices that require internal file storage (since FAT-x is or has become public domain), and (b) to prevent file access by local users without logging onto the system (ie without booting up the full NT-based OS itself) and when the OS is up and running you have another layer of file access defined by user and group rights.

None of that is possible with FAT-32 (which is a weakness in corporate, institutional or military environments) but none of that is necessary for the home or SOHO user. Unless you can rely on your local IT department for computer maintenance (because they are likely to have access to expensive third-party NTFS repair utilities) then it's far easier, and cheaper, for the novice to repair FAT-32 volumes. The argument that NTFS is more reliable or robust than FAT32 is pure fiction. I have probably more than 1000 years of collective FAT-32 hard-drive experience, and have never lost data due to soft-errors caused by logical corruption. FAT32 is faster given it's lack of journaling and other overhead, and FAT32 is also more secure from a virus, trojan and root-kit perspective because (a) it's harder to hide files and prevent AV access to files on a FAT32 volume, and (b) NTFS alternate data streams are vulnerable to malware usage (and naturally, FAT-32 does not have this rather useless "feature" in the first place).

I use NTFS on some of my home PCs and it's very good.

And how exactly do you score, rate or define the "goodness" of NTFS?

Your statement is similar to how many people rate their AV software. They come to the conclusion that the particular AV software they use is "good" because it rarely, or ever, alerts them to the presence of malware. Whether that's because they've never exposed their system to malware, or because they have malware that is undetectable, is a conundrum they never consider.

Naturally NTFS is a _functional_ file system. But compared to FAT32, it's an overly complex file system for many of the environments and situations it's used for. And there are always downsides to unneeded complexity.

My motto for Microsoft: If it works, it's not complicated enough.

[North of Watford]

98Guy, I share your concerns about the NTFS file system, especially about the ability to hide stuff in the alternative data streams.

That enough people have sussed out how to write utilities such as NTFS drivers for Win98 must indicate that MS didn't make it complicated enough. Possibly the reason for the development of HFS, their newest way to confuse their enemy (the customer)!

That said however, I do find it necessary to use NTFS for a few jobs. I recently had to write some video files to a Video DVD for someone without a PC. Whilst it can be done in Win98 (with some difficulty in preparing the files), the simplest programs to use seem to be written for Win2K upwards, and they insist on NTFS for the freedom of creating large files. I initially ran one such program on a Win2K box with only FAT32 formatted disks, and it reliably fell over every time.

The Paragon driver noted above is still limited to a 4GB maximum file size.

This then leads me to question whether the next big leap forward for Win98 is to overcome the 4GB file size within both Win98 itself, and within a suitable file system. Perhaps it should adopt the Ext2/3 as native? Could it be done? Does someone have the technical understanding to do so?