Re: Archiving

On Fri, 18 Jul 2008 02:08:24 GMT,
jellybean stonerfish <stonerfish@xxxxxxxxxxxxx> wrote:

Hmmmm... I wonder if gnu tar could be fiddled with to add an option to
compress each file first, then wrap the compressed/encrypted files in the
tar archive. This could mean losing only one file, instead of the whole
archive if there is a flipped bit.

Personally I'd take a different tar (perhaps `bsd tar', or `star' ) if
I wanted to fiddle, as gnu tar for all its ubiquitousness (hah[1]) is
a source of interesting compatability problems itself. But you don't
really need to touch the source: Simply gzip'ing the files first, then
tar, would work, though it'd be inconvenient without at least some
wrapper shell script.

OTOH I wouldn't go there in the first place. Such a solution exists
already, and is called 'zip', or 'arj', or 'lha', or 'rar', or even any
of a bunch of obscure ones (anybody remember 'uc2', or 'hap/pah'?).
I recall that zip stores its file index at the end of the archive so
that may not be the best choice, though it does come with some index
reconstructing utility (at least the dos utility pkzip did).

In this context (compressing first, then) adding redundancy is an
interesting option and one example is the `parchive' (also apparently
a fork called `par2') software that will let you create `parity files'
with which missing chunks of the original can be reconstructed up to
some (choosable) limit. Integrating this into an archiver is fairly
obvious as a goal and it seems rar has a (ISTR not very efficient)
implementation of this.


[1] People in the gnu world like to believe that and advocate it saying
all the cool kids are doing it, but that doesn't make it strictly true.
Or maybe there just are a lot of uncool people around. Who knows?

--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
This message was originally posted on Usenet in plain text.
Any other representation, additions, or changes do not have my
consent and may be a violation of international copyright law.
.