Date   : Thu, 09 Dec 2010 12:26:09 -0600
From   : jules.richardson99@... (Jules Richardson)
Subject: FS: Master 512 - New and Unused,

J.G.Harston wrote:
> Jules Richardson wrote:
>> I'm sifting through stuff... slowly. Unfortunately a lot of my disk   
>> images are corrupt
> 
> Hmm. I've had experience pulling stuff out of corrupt disks/images,
> usually coping with bitrot from old disks. I might be able to help.

Thanks - but the nature of the corruption is completely missing data, rather 
simply garbled bits; there's no way to reconstruct valid data from what I have 
here.

It was my own darn fault - I sanity-checked a random sample of the images 
prior to moving over here and at a glance they looked OK; file headers were 
good, sector counts according to track headers appeared good etc., and there 
was visible sector data in there, so I didn't decode further. The problem is, 
random chunks are missing from the files - and not aligned with sector or 
track boundaries, either (e.g. I might see the first half of sector 5 on track 
10 in an image file, say, followed immediately by the end half of sector 8 on 
track 11).


The nature of the corruption suggests that it probably didn't happen at 
creation time - I've got the source for the creating util and it just wouldn't 
write 'incomplete' junk like that, even if thrown a severely-damaged disk; 
it'd either store each sector's data intact or flag a sector as unreadable. 
Assuming, of course, that there wasn't something weird like a memory fault at 
play...

At the moment I'm tentatively pointing my finger at something in the data 
transfer process between the 'reading' machine and my main PC as being the 
most likely culprit - either the DOS TCP/IP stack or FTP program that I used, 
or the network hardware. When I eventually do go back to the UK I can 
sanity-check file sizes on the reading machine against what I have here and if 
that's the problem transfer data by other means.

cheers

Jules