Let's describe the format of the C64 Collections maintained through SceneBase. We're going to refer to them as SceneBase archives or 'SB archives' for short. In fact we use SB to "tag" the file names of all SB archive files.
Collection File Naming
For every collection we process, at least two and maybe four files will be produced. We use the same name that the original collection was released with as a base and then use short tags to identify the SceneBase version. With few exceptions for portability, the file name base should be identical with the original... for example, spaces are exchanged with underscores. As a live example, "C64_DiskImagesBox01_DMC.zip" is the original name of a collection. The resulting SB archives will be:
C64_DiskImagesBox01_DMC.SB.7z
C64_DiskImagesBox01_DMC.SB-info.7z
C64_DiskImagesBox01_DMC.SB-nidx.7z (optional)
C64_DiskImagesBox01_DMC.SB-dupe.7z (optional)
Let's explain the different files, per "tag" used:
- SB: The main file which will contain only media images inside a single directory named identical with the collection name (in this case, "C64_DiskImagesBox01_DMC").
SB-info: A file with various metadata that describes the contents of the media images. Currently we produce a .txt (ascii) directory listing; a .csv (comma separated values) representation of the directories with MD5 hashes per file on the image; and one .png per image showing a screenshot of the native (petscii) directory listing. We also provide a file of the form "rename.[collection name]" that maps the SB media image names to their previous form. In addition, if the original collection contained any metadata content of its own, we keep that data here in a directory named "original". We keep metadata in its own file to provide for persons interesting in conducting searches for specific native file names without requiring a transfer of multiple Gb of media images.- SB-nidx: A file with content from the original collection that is neither media images nor metadata. Examples could be photos of collections, scanned documents (like disk covers), etc. Any files that aren't well preserved replicas of actual media - like individual .prg files - are also shifted to the nidx (no-index) file. Not all SB archives have a corresponding nidx file.
- SB-dupe: A file with images that are bit-equivalent duplicates of disks within the current collection or across all collections. This is a future project and we have not yet started to separate duplicate images from any collection.
In a few cases we've split an original collection release into multiple files with the goal of producing more easily downloaded archives. For those collections, we append ".a", ".b", ".c", and so on immediately after the collection's base name and before the SceneBase archive tag. For example, "C64_DiskImagesBox01_DMC.a.SB.7z" and "C64_DiskImagesBox01_DMC.b.SB.7z" would the the two main files of a multipart collection.
Collection File Compression
Note that the main media image files (tagged SB) are solid archives. However the metadata files (tagged SB-info) are not solid archives, idea being to sacrifice relatively small amount of size in exchange that researchers can quickly extract individual files as fast as possible while in 'browse mode'.
Media Image File Naming
The media image files contained in the SB archive main files are of the form, "sb######.xxx" where "######" is a six character id used to uniquely name each media image and "xxx" is whatever format extension applies (d64, t64, d81, etc). The .png screenshots in the metadata (SB-info) file that correspond to each media image are named using the same form, aside from, of course, using a .png extension.
All Metadata Single File Download
Check the right bar for a link or links to a single file containing the metadata for all c64 collections that have been processed to date. The 'all metadata' file name is date stamped so future updates are discernable.