[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: My laser 128

Antoine Vignau wrote:

On 10 jan, 20:52, Michael J. Mahon <mjma...@aol.com> wrote:


D Finnigan <dog_...@macgui.com> wrote:


Bill Hicks wrote:


On Jan 10, 1:07�am, Michael J. Mahon <mjma...@aol.com> wrote:



Exactly. There was a short window when "single sided" disks were actually
single sided.



And Apple's method of formatting is more tolerant of isolated surface
imperfections than others.



-michael - NadaNet 3.1 and AppleCrate II:http://home.comcast.net/~mjmahon



When you init a disk does dos do any verification?



From reading some documentation on ProDOS, I got the impression that the
ProDOS Filer utility does make an attempt at marking out bad blocks so that
they won't be used to store data.


Nope. A disk either formats or it doesn't--there is no mechanism for
altering the track capacity.




Correct for track capacity.
But what prevents Filer from reading blocks after a format and mark
out bad blocks? That means that Filer won't stop on I/O while
formatting though.



When formatting track 0, if it does not read back correctly, the track will
be reformatted in a different rotational phase, which can bypass isolated
media faults if they fall in sync areas.

After a certain number of retries, it gives up.




For my knowledge, from where did you get that information, Michael?


I don't have a specific reference--that was just from memory.

I just checked a couple of places I might have read it, but came
up empty.  In fact, I haven't found any specific information on
the ProDOS Filer formatting algorithm.

It's even possible that I was (mis)remembering the DOS 3.3 formatting
algorithm, which, as you say, isn't satisfied with a track until all
16 sectors can be read successfully.

In reviewing the DOS 3.3 algorithm, I also see that the format write
does not start at a random location on each retry, but starts just
after sector 15.  As the number of self-sync nibbles is decreased on
each retry, this will result in precessing the sectors around the disk,
but the process is quite orderly, not random.

The net effect is still to migrate sectors and gaps around the
disk surface, which provides some tolerance of surface imperfections.

I had remembered that it did this only for track 0, but it appears to
repeat this as part of its error recovery if formatting fails on any
later track.

-michael

NadaNet 3.1 for Apple II parallel computing!
Home page:  http://home.comcast.net/~mjmahon/

"The wastebasket is our most important design
tool--and it's seriously underused."