The MMB format is very limited in what it can handle; basically just 200K disks. I can understand the attraction; simple is good!
However it makes it hard to extract floppies from double density disks. 320 or 360K of data just doesn't fit into an SSD.
So what I've written is a bit of a kludge. It's a program that will load tracks (OSWORD &7F) into memory, and then save them as files on an SSD (two tracks per file). Of course this won't fit onto one disk, so tracks 0->39 go onto SSD1, and 40->79 go onto SSD2. Then the resulting SSD images can be extracted on another machine and the sub-files extracted and concattenated together.
eg on Unix using my perl routines:
Code:
#!/bin/sh
F=/media/0C0F-C2C0/BEEB.MMB
rm -rf part1 part2
beeb dgetfile -f $F 1 part1
beeb dgetfile -f $F 3 part2
cat part?/PART?? > disk.dsk
"disk.dsk" is now a copy of the original disk.
Of course, the common emulators can't (currently) handle this format, but it's a good start to preserving the raw images. Or files can be extracted from them and copied to SSDs.
If there's any interest in this I can make the code a little more "resillient". I've only been able to test on my Beeb with TurboMMC and Solidisk DDFS but the test single density disk I was using gave an identical image when extracted this way or by using IFEL's IDTOM imager ROM.
(To try and make it reliable whenever I switch between DFS and MMC-DFS then I swap pages &D->PAGE so the DFS only sees DFS pages and MMC only sees MMC pages; I also disable MMC via poking into it's slot in &2A1 to stop it stealing OSWORD &7F calls, then re-enable it).
