Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!iuvax!maytag!watstat!dmurdoch
From: dmurdoch@watstat.uwaterloo.ca (Duncan Murdoch)
Newsgroups: comp.sys.ibm.pc
Subject: Re: DOS 32 Meg limit
Message-ID: <1990May29.022140.11755@maytag.waterloo.edu>
Date: 29 May 90 02:21:40 GMT
References: <2381@zipeecs.umich.edu> <446@accucx.cc.ruu.nl> <4915@plains.UUCP>
Sender: daemon@maytag.waterloo.edu (Admin)
Organization: University of Waterloo
Lines: 28

In article <4915@plains.UUCP> kperson@plains.UUCP (Kerry Person) writes:
>In article <446@accucx.cc.ruu.nl> aceverj@accucx.UUCP (Jaap Verhage) writes:
>>>
>>Yes, all of this may be true, but the 32 Mb limit has
>>disappeared with DOS 4.0. I don't know what the limit is now,
>>but it's certainly more than it was.
>
>
>I thought it was the square of the previous limit, since the address now has
>twice as many digits.  Or am I all wet?

It's not quite the square - other things get in the way.  It used to be that
64K of sectors limited your disk to 32 Megs.  Now there's effectively no
limit on the number of sectors, but there's still a limit of 64K (minus some
special codes) on the number of clusters.  This means that you can't go
bigger than 128 Meg with 2K clusters, 256 Meg with 4K, etc.  I don't know
the largest cluster size DOS 4 can handle, but I doubt you'd want the 
32 megabyte clusters on a disk with (64K)**2 sectors.  (On the other hand,
on a 2 terabyte disk, it might not matter.)

By the way, does anyone know a reliable test for whether a particular copy
of DOS can handle/requires the new long sector numbers in absolute read and
write?  Zenith DOS 3.30 Plus can and does on big disks, but just about every
other 3.30 (which all look identical if you just ask for a version number)
can't.  I think it's safe to say that any version 3.31 or higher can use
the new calls; is this true?

Duncan Murdoch