I see here that a tar is used to merge several files into one. So why do I see so often one zip file in a tar?
example : archive.zip.tar
Tar doesn't compress* - it's 'tape archive' and exists purely to serialise data - traditionally to a tape device.
It serialises a directory structure, including permissions and file contents.
gzip doesn't preserve structure, and just compresses on a 'file' bases - hence why you tend to get
.tar.gz - it's a serialised structure that has then been compressed.
I don't know why you'd get
.zip.tar though, given that
zip does support directory structures, permissions and compression. It seems largely nonsensical to do so.
* newer versions of
tar include the
z flag to compress - this wasn't part of the original spec. So you'd
tar cvf - ./path_to_tar | gzip -c > file.tar.gz