Consider storing the original images as well

[msikma]

Hello, I think the data that's been preserved in this repo is very interesting, and I applaud the efforts to rescue this data from the original disks. I wanted to make a suggestion that it might be worth preserving the original disk images as well, or potentially storing the original files in zip archives.

The main reason is that files have file creation/modification metadata, even on floppy disks, and this data ends up being lost when storing the files in a git repo. Git repos aren't designed to store information at the file metadata level, with the files being dynamically recreated based on e.g. switching branches. For old binaries and source code like this it can be quite useful to keep the file timestamps, for determining exactly when a particular file was worked at, or when exactly a particular binary was created relative to others, and so on.

Even if it's difficult to preserve disk images, zip files do store file creation/modification metadata as well and will properly set them on extraction, so that's potentially an option as well.

[sdstrowes]

+1 (though sticking binary blobs into the repo is also not ideal; sticking them in block storage somewhere would be better)

Also: the readme notes that there were 1,000 disks, of which 500 were captured and the others "could not be read due to degradation of the media". DSDD disks are surprisingly durable and I wouldn't be surprised if a greaseweazle plus a couple of different drives and the patience to clean the surface of some of the disks themselves wouldn't get more of them. I've never seen a 50% failure rate. Depending on the location, I'd be interested in taking a crack at this.