Executor 2.0/executorv2.0.iso/pc/dos/extra/docs/maillist/text/archive.96/text3970.txt

Text File | 1996-07-25 | 5.9 KB | 118 lines

In-reply-to: [e-mail]'s message of Thu, 28 Mar 1996 13:06:24 GMT
To: [e-mail]
X-MailNews-Gateway: From newsgroup comp.emulators.mac.executor
Sender: [e-mail]
Precedence: bulk

-----BEGIN PGP SIGNED MESSAGE-----

In article <[e-mail]> [e-mail] (Francois Roy) writes:
> Rob Heath ([e-mail]) wrote:
> : is there any program out there that will read 800k mac floppy disks?
>
> As far as I know, there isn't.  From what I've heard, the 800k floppys
> are written with a different speed of rotation depending on whether you
> are writing on the outer or the inner tracks.  The drives on PCs all
> operate at a constant rotational speed.

It is a common trick (used by harddrive manufacturers) to increase the
data density on the outer tracks, but I do not know whether the Mac
actually makes use of this for its 800k floppy disks.

But the main problem with trying to use 800k floppy disks on a PC is
different from that (I try to explain just the basic concepts; if you
do not like the simplifications that I make, either get hold of good
documentation on these topics or ask me and I will try to look it up):

The data on each track of a the floppy disk is basically a cyclical
stream of "one" and "zero" bits. There is no way, that the hardware
could tell where the data starts and where it ends. Thus this
information needs to be somehow encoded into the data stream; a very
simplistic approach could work like this:

 1) expand every byte ( = 8 bits) of data to be 9 bits long
 2) for regular user-supplied data, that is to be stored on the disk,
    set the 9th bit to "one"
 3) for synchronisation information, set the 9th bit to "zero"
 4) mark the beginning of each data block, that contains just
    synchronisation data, but nine consecutive "zero" bits.

This way we would waste a little more than 11% of disk capacity, but
we could now tell where to start looking for data. Unfortunately, this
will fail if the user data contains long blocks of 0xFF bytes (all
bits set). Small deviations in rotational velocity would cause us to
get out of phase and we would then read bogus data.

Therefore, encoding mechanisms have been designed which guarantee,
that there will be no more than a certain amount of consecutive bits
with the same value. This enables us to obtain clock information and
re-synchronize on the data stream while reading the data. Besides, it
offers a new improved technique for marking the beginning of
synchronisation information. Now, there are sequences of bits, that
are definitly "illegal" with regard to our encoding scheme; these
sequences are used to mark "out-of-band" data.

As computer technology improved, different encoding schemes were used;
some of the more important schemes where named FM, GCR, MFM, and
RLL. In early days, the computer's CPU would actively do the
encoding/decoding. Thus it did not really matter, if a new encoding
scheme was developed; as long as the CPU was sufficiently fast, all
you had to do was upgrading your software. This was the case for the
Commodore C64 and for the Apple II. Later, some dedicated hardware was
designed that assisted the CPU while still retaining flexibility and
allowing for different encoding schemes. As far as I can tell this is
the case with the Commodore Amiga and the Apple MacIntosh. At about
the same time, MFM encoding became very popular and developed into the
dominating encoding scheme. Thus chip designers, developed a "floppy
disk controller" that would do *all* of the encoding/decoding. This is
what is done in current PC's.

The PC has always used MFM encoding (even though RLL encoding would
bring a 50% improvement in data capacity), but the data density has
increased considerably from single sided, 40 tracks, 9 sectors (this
gives 180kB) to double sided, 80 tracks, 18 sectors (this gives 1.44
MB). I believe there have been some other formats that stored even
less data and there also is a format that stores 2.88 MB (while using
36 sectors), but I think these have never been really popular.

The Apple MacIntosh used to have its own proprietary format. I do not
know much about the encoding scheme other than one reference that I
found; it states something about "6-out-of-2" coding. Back in the
days, when "double density" disks (720kB) where popular, the Apple
MacIntosh had the advantage of storing about 10% more data (800kB) on
the same medium. But when "high density" disks (1.44MB) became
commonly available, Apple had to decide whether they wanted to come up
with an improved proprietary format or if they would go with the
"industry standard". Fortunately, they decided for the latter. This
has the disadvantage of storing less data than is physically possible
(I believe, advanced encoding schemes can store at least up to 1.8MB
user data on a regular "high density" disk), but finally there is just
one standardized format, which allows for easy exchange of media.

Computers such as the Commodore Amiga or the Apple MacIntosh still
have the ability of handling arbitrary encoding schemes, if the CPU is
programmed to recognize this format, but the PC never had this
ability. There are tricks that can be used in order to force the
"floppy controller" into accessing the raw data stream, but this is
usually very unreliable and you will never read the same data
twice. Thus, with sophisticated software it might just be possible to
make the PC read 800kB floppy disks, but I would not trust it to
reliably write these disks.

I hope, I cleared up some of the confusion,


Markus

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Processed by Mailcrypt 3.3, an Emacs/PGP interface

iQCVAgUBMV+X7xqJqDLErwMxAQFzPgP/YifZAIPZJl5pA7tatIW8rB/rfEx4VhMz
6D7u+oQPpTUswmqmbgyFe9Y6JsfKNvzryqt+WCb1Sqeh06bXi57GcB5sLAM8PfRk
6Cu6HtVzgAXKEraCOOGNjod/ApO8vLAal02nAWivwjMCjq8KiWdMgCjp2EtJj710
vf9terdgnTM=
=V7ai
-----END PGP SIGNATURE-----