Re: Hardware musings

[russotto@grace.speakeasy.net (Matthew Russotto)]

In article <Ap-dnZrH5JYZ76DYnZ2dnUVZ_u-dnZ2d@comcast.com>,
Michael J. Mahon <mjmahon@aol.com> wrote:
>Matthew Russotto wrote:
>> In article <aJKdnTNJ_tZU6qXYnZ2dnUVZ_o2dnZ2d@comcast.com>,
>> Michael J. Mahon <mjmahon@aol.com> wrote:
>> 
>> 
>>>There is the potential for writing a track with ambiguous nibble sync,
>>>in the sense that it could be read with nibbles synced in more than one
>>>way.  I don't know of any program that uses such a technique, but it is
>>>possible.  
>> 
>> 
>> I ran into one such, the driver disk for a speech recognizer.  It
>> didn't seem to have any self-sync nybbles.  It was slow as heck
>> running its modified DOS (which I presumed, but didn't prove, tried
>> every possibility until sectors started reading).  Since it used a
>> normal DOS filesystem, I was able to copy it file by file to a regular
>> disk, using a cassette tape for an intermediary.
>
>A very creative solution!
>
>That would be a natural for a "swap controller" that swaps
>between the protected RWTS for reading and a normal RWTS
>for writing.

Yep, that's what I tried, but it didn't work.  I could read sometimes, but not
at all reliably.  My best guess is that they'd modified the code
calling the RWTS to change the timing of re-tries in ways which
would semi-reliably change the starting bit to something else on the
retry.  But that's when I had the cassette-tape idea and I didn't
investigate further -- working with that disk was painful due to its
speed.
-- 
  There's no such thing as a free lunch, but certain accounting practices can
  result in a fully-depreciated one.