🚀 go-pugleaf

In article <1990Jun11.114856.1095@bbt.se>, pgd@bbt.se (P.Garbha) writes:
> In article <1317@chinacat.Unicom.COM> woody@chinacat.Unicom.COM (Woodrow Baker) writes:
> 
> >
> >There are many and varied problems with DOS 4.0.  Many of my coworkers
> >have tried it, and invariable have had to go back to dos 3.x.
> >
> 
> What are the problems with DOS 4.01? we just starting to use it because
> of windows 3.0. DOS 3.3 is not the greatest, for the single fact that
> it does not allow hard-disks greater than 32M, and i strongly dislike having
> 5 or 10 partitions on one disk.

O.K. for a bit of background: MS-DOS version 1.25 was a very simple OS.
It only supported floppies, and as such, it used a 12 bit fat table.  
MS came out with DOS 2.1 or 2.0.  This supported hard disks.  To do so, they
expanded the FAT to 16 bits.  Now, since the number of copies of 1.25 was
very small, and most of the programs were rather simple in nature, that 
caused only minor problems.  Now along comes IBM and re-writes Dos 3.3
(AT this point:  I have heard 2 stories, 1. IBM wrote it and licensed it
back to MS, 2. DR wrote it and licensed it to MS, I don't know which
is correct).  The major advance for DOS 4.x was the support for larger
hard disks.  Who ever wrote it, decided that the way to do this, was to
change the FAT table size.  With 15-35 MILLION copies of DOS floating around,
versus 10-20 thousand, the impact of doing this was much worse.  I remained
truely puzzled as to why DOS 4.x seems to have problems.  Our commercial
software package would not work.  One of my clients is a CPA firm that
uses a commercial software package.  They had DOS 3.x.  Put Dos 4.x on
a machine, and when they got an upgrade, the software failed.  Not only
failed on various machines, but also failed on DOS 3.3 (the install
program had munged the DOS 3.3 disk).  After 4 or 5 shipments of new
software, and various package configurations (to keep magnetic things)
from messing with the software, they casualy mentioned to the software
maker that they were using 4.01.  They were advised to remove it immediatly.
they did, the problems went away.  They were also using wordstar.  Files
would flat dissapear from the floppies etc.etc.  Our software product
was written with Lattice 2.0.  Due to lots of compiler dependent code, it
is not feasable to migrate it to another 'c' compiler.  It won't run reliabley
under dos 4.x.  I think, after looking at some programming documentation, that
I have figured out why some of the problems exist.  Under Dos 4.0, certain
function calls were modified.  Specificaly the CX register is now used for
disk functions where it was not used before.  If it has a value in it, I think
a -1, AND you are running under DOS 4.x, Dos expects one of the other registers
(I think DX) to be pointing to a parameter block in memory.  If cx is NOT
-1 then DX holds something else.  Since, under 3.3, CX was not used, software
often left the contents of CX as it was, or actualy save temporary values
in it.  If it happens to be -1 AND the code was written under 3.3, then
the DX (or whatever it is) is not going to be pointing to what it should,
and boom.  Anyway, all of the problems that I have seen when people move
to DOS 4.x vanish when the move is made back to dos 3.3.

I think that the culprit is the extensions to the disk services.  I only
briefly glanced at one of the function call specs for DOS 4.x, so I may
have some of my registers off, but the general operation, I think is correct.

Cheers
Woody