> Oh dear - if you need to ask this in here, it is a fair indication
David, that is exactly what I have expected as a response to my
entry, just nothing.

 > that you aren't ready for it.
I'm here in Australia 6 years (I'm from Praque), but still can
remember a saying from my home country ..... Think twice, then
say...


 > However, if you really want to screw up your hard drive, then get
 > "Undocumented DOS", found in any good bookstore. You'll need to be
 > fooling around with INT 25H and INT 26H for direct disk access (it
 > is no longer a simple call - you need slightly different methods
 > for 12-bit, 16-bit and DOS 3.1+ partition types), as well as know
 > how to page in and read the FAT, update the FAT and flush it and
 > tell DOS that it has changed.
However David, I really don't want to send away my HD or my FD.
My, let me say test module is working. I'm  very familiar with
the FAT structure, can of course recognise between 12 and 16 bit
fat table entry and know how to ship 12 bit FAT entries for next
actions. Moreover this my test module is able to say how bad or
good is the cluster fragmentation on HD. I can establish my own
buffer for DTA ( get int21 fce 2f and set int 21 fce 1A ), that
means I can read  the attribute without fce 4300 and of course I
have the start cluster number available as well.   But this is
not nothing special. The reason why I have put my update in this
conference call is that more people means more ideas and I think
that any discussion about any problem can be always useful even
with no top professional people, but only in case if non of them
thinks I'm the best and I know everything.


 > There's nothing overwhelmingly difficult in doing it, but it's one
 > of those things that needs to be *thoroughly* tested since a very
 > minor slip causes one major headache. And it'll stop working on
 > many types of disks including networks, compressed drives, WORMs,
 > CDROMs etc. where the BIOS and/or device driver dos the block I/O
 > via different methods, and just plain won't work for any
 > protected-mode environment which offers a simulated MSDOS session.
 > The situation where direct access works these days are becoming
 > fewer and fewer, so the investment required in writing and testing
 > code isn't really worth it (imho).
In general you are right, but the path always exist. Just to know
more about the system/subsystem. Not all of the software
manufacturers are willing to say what is after the door. My
testmodule works fine even under OS/2 dos box ( but they are not
sufficiently compatible  with the crit. err. handler int 24 -what
I have found so far ). Now I like to a bit discover how
internally works the HPFS ( High performance file system ) in
OS/2. As you know under OS/2 you can have 'classic' FAT or HPFS.
Okay next piece to play with. You should understand that all that
is my hobby only, because I like it. ( My job is a mainframe MVS
system engineer mainly using assembler).

Okay David, thank you for your update anyway and if you like why
not have a beer (from my home production) and a chat about
'lovely computing'.

Regards, Vladimir


---