Date   : Mon, 27 Apr 2009 14:49:30 +0100
From   : alun.rhys.jones@... (Rhys Jones)
Subject: Domesday Preservation

On 27 Apr 2009, at 13:45, Jules Richardson wrote:

> Rhys Jones wrote:
>> The setup was to follow the Camileon process of hooking up the LV-ROM
>> to a Linux box via scsi, and to parse the data to save each file from
>> the disks.
>
> "was" as in "not done it yet"?
>
> You have me curious, as for the player firmware I've seen, the units  
> don't
> quite support full SCSI CCS, and Linux barfs in trying to initially  
> talk to
> them (and I've not found a way of working around this so that read/ 
> write
> commands can still be sent to a device that the kernel thinks is  
> 'broken').
>
> It's entirely possible that Philips released later firmware which  
> was a bit
> more functional, however (in which case a dump of the relevant ROM  
> would be
> most useful! :-)
>
> I can't remember what Camileon did now - I'd have to dig through my  
> email
> archives. I know Adrian P. dragged the data across via the RS232  
> interface
> which took *many* hours.
>
> cheers
>
> Jules
>

The setup took a while to get right - the first Adaptec 2940UW card I  
tried did not detect anything, but then one bought off ebay (a Compaq  
branded 2940UW card) showed up the LV-ROM as drive D: during the boot  
process, but would complain that the device had no name etc. The  
Camileon documentation said that they needed linux 2.4.7 to get the LV- 
ROM working, and I iterated through a few linux distros to end with  
Redhat 7.1 finding the drive under /dev/sg0.

Interestingly, the original Camileon project docs and source seems to  
be only available on the UK Web Archive[1]. The code made available on  
there worked fine in pulling the data off the disks in chunks and then  
re-combining them into a disk image.

Running 'strings laserDiscData.ldi | grep SearchTerm' on the image  
file gives the raw text, but makes for a poor domesday client...

The RS232 code was much easier - a small python program talking to / 
dev/ttyS0 would send the FxxxxxR command to go to a frame, check that  
the LV-ROM got there, and then run mplayer to grab a png of the image.  
I found that the LV-ROM would over-skip some frames, so needed the  
explicit check and re-try to be sure of the right one.

Hwyl,
Rhys

[1] http://www.webarchive.org.uk/ukwa/target/99941/repnet/casestudy.html