zip (was: Re: Disk archival techniques)

Is there a reason to settle on a single format rather than two? Store both a zip file and a gzipped tar file. (Or, for that matter, why only keep a single copy of an archive)? A cheap ($100) hard drive holds about 312,000 uncompressed DS/SD 5.25" floppy images. Next year's drive will hold 500K. I'd guess that's enough for every CP/M software package ever published (plus source code) with room left over for every paper tape currently in existence and a few VMS distributions. I usually store a zip file, a .tar.bz2 file and a bz2 compressed Floppy/CD image where applicable. Copy protected CDs are a problem because I don't know of an open format, so I'm currently using VirtualDrive to make backup images of them. Compressed ISOs work fine for unprotected disks. My current (home) archiving technique is to have two networked RAID-1 setups. One is old technology (in the case a pair of 30GB drives). The other is newer tech (a pair of 100GB drives). I've just ordered a pair of 400G drives to replace the 30G drives. When they arrive, I'll archive the 30G drives to tape, I'll copy everything from the 100G on to the new drives. The the 100G drives become read only, and the 400G drives become the active archive. The 30G drives go into sealed CO2 storage. The active archive always contains a copy of the older archives and gets backed up regularly. When 1.5TB drives hit the market, the process repeats itself (probably with new hardware, since there might not be 1.5TB ATA drives). If an archive format conversion is required (as was for .ar[ck] to .zip or .tar.gz to .tar.bz2), it's easy enough for a script to handle the job. Since I also keep source code to gtar, gzip, and an unzip on the drives and executables in for every machine and OS I've ever used, things should be recoverable, if kept up to date. Eric On 5/17/05, Al Kossow <aek at bitsavers.org> wrote: