moss@BRL-VLD.ARPA (10/01/84)
From: "Gary S. Moss (AMXBR-VLD-V)" <moss@BRL-VLD.ARPA> Remember, this discussion started out with VMS style multiple file versions and an unremove command, some of your ideas sound neat, and some a little far out. There is a capability that exists in the way of public domain software with Doug Kington's 'Multiple Device Queueing System (MDQS)' to run batch jobs on UNIX which will restart themselves after a crash (they do start over from the beginning, but heck it's free). All I am saying is RCS and SCCS are much more advanced than 'automagic' multiple versions and they don't step on people's toes who don't choose to use them. -- Moss.
fouts@ames-nas-gw.arpa (10/02/84)
From: Martin Fouts <fouts@ames-nas-gw.arpa> Actually, all of the things on my wish list of been implemented in one system or another, most before 1975, although a lot of the early emphasis was on their use in batch systems. We are looking at MDQS, but still have to solve the problem of checkpointing jobs, which is just one of the mundane problems related to the way compute intensive numerical calcuations are done now. I agree, from the point of view of data compression that RCS / SCCS provide a better solution than version number schemes. But a version number system which stored diffs rather than whole files would have the same advantages in this context. I also agree that version numbering should be an option, but I still maintain that situations arrive in which keeping (in some sense) all versions of a file make sense. ----------