ImageDisk/Site update posted.

Allison ajp166 at bellatlantic.net
Thu Aug 4 13:15:58 CDT 2005


>
>Subject: Re: ImageDisk/Site update posted.
>   From: Dave Dunfield <dave04a at dunfield.com>
>   Date: Thu, 04 Aug 2005 12:01:10 -0400
>     To: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at classiccmp.org>
>
>At 10:38 AM 8/4/05 -0500, you wrote:
>>Dave Dunfield wrote:
>>
>>> Hi Guys,
>>> 
>>> I have posted the update to ImageDisk which allows it to work
>>> with Head/Cylinder values on a per-sector basis if needed.
>>> I have also posted a number of additional ImageDisk format
>>> system disk archives.
>>
>>Dave, from my quick read of your description of the IMD format, there is meta 
>>information that I think would be valuable to store.  When a disk is imaged, the 
>>captured file should indicate which tracks/sectors had problems being read. 
>>That would allow automated merging of two different attempts to capture a given 
>>disk (I haven't used your merge program, but my understanding was that one had 
>>to manually specify which tracks to take from each image.  Futher, when 
>>archiving one often is processing a large batch of disks.  Keeping track of the 
>>problematic sectors manually and making sure that information is associated with 
>>the disk is challenging.  Some months later when going back to a pile of disk 
>>images, it would be nice to know which sectors are suspect.
>
>Jim,
>
>I *DO* keep track of which sectors could not be read. If you look at the SECTOR
>DATA RECORD, you will see that it consists basically of a flag byte followed by
>the sector data. The flag bytes are:
>
> 00 = Sector unavailable (no data follows)
>      This means that the sector could not be read from the original master.
>
> 01 = Normal sector data (sector-size data bytes follow)
>      Self explainatory.
>
> 02 = Compressed sector data (1 byte follows)
>      This means that the entire sector was read and found to have all bytes
>      containing the same data value - IMD will automatically expand this sector-size
>      occuranges of the data value - it's primary purpose is to reduce the storage
>      required by unused areas of the disk.
>
>My analyze/merge utility will automatically extract 01 or 02 sectors from one image to
>fill in 00 "holes" in another. It will also warn you if 01 or 02 sectors occuring on
>both disks don't logically match (in which case you probably shouldn't be merging them).
>
>At this time I DO NOT keep track of the occurance of soft/recoverable read errors. In
>my experience these have more to do with the particular PC/FDC/DRIVE combinarion and/or
>individual disk condition, and I do not believe the added complexity/size of collecting
>and recording such details will be worth it...
>
>ImageDisk will display sector read errors, and unavailable sectors as it records the
>disk, and analyze/merge will display a summary of the Unavilable, NormalData and Compressed
>sectors occuring within an image, as well as a detailed track by track summary if requested.
>For me this gives me pretty much everything I can actually make use of.

Dave,

Read diagnostic command will read sectors with bad CRC or other problems.
That may help with unreadable sectors.

Keep cranking, you have a very useful tool there.

Allison




More information about the cctalk mailing list