TAN files have been designed to be shared and linked, just like any network of files. Most often, TAN files will be created and distributed as collections, not single files.
One way to distribute a collection is by making it available as a repository via Git or some other version control software (VCS). This approach has many advantages. The files become available to anyone who wants them, and the editorial history is preserved. VCS features and tools are extremely fast and useful, and they allow users to modify TAN collections without impacting the original source.
Collections may also be distributed through shared syncing services (e.g., Drive,
Box, or Dropbox), or put on a Web server. In the latter case, it may be difficult for
users to browse or download wholesale. In that case, you may wish to expose the
collection as a compressed ZIP archive. This saves on your server's bandwidth, and it
still exposes the files for XML processing. But a ZIP archive is not suitable for
linking from one TAN file to another, nor is it appropriate as a <master-location>
. Unpacking
a compressed file requires writing to the disk, which is considered to be a security
risk during validation, and so is disallowed. Such zipped archives are good ways to
distribute a collection, but they should not be used as a primary repository or a
master location.