1) Incremental backups with only a small 'catalog' as the reference instead of having to have the full backup available
2) Ease of splitting the backup set over multiple devices with the ability to unmount the device in the middle of a backup once it is full to mount a new volume.
These things can be done with tar, but not without some extra work.
For this to be an accurate comparison, the word processor would have to support only one formatting option -- dar only has one form of compression built in, does not seem to support writing to stdout, and was last updated in 2006, so the advantages of a pipeline setup and having the ability to choose alternative tools as they become available outweigh the limited feature set over more traditional tools.
However, there is this (from the man page):
"Because, even when using dar to write a slice on a tape,
you will loose all the interest of another feature of dar
which is its ability to directly access any of the
archived files, even when compression is used."
which can be tremendously useful, but isn't compelling enough to warrant using it.
Ah, I missed that. Now that I look closer at the "Current Status", you're right. That is a not-very-straight-forward way to present releases and schedules.
I didn't say it was an exact analogue I said the point was the same - simplifying the process using a gui (some think those things are mutually exclusive!). Nor did I even imply that dar is a good solution.
Now if you'd said what's the point of it versus ark, or more likely, rsync or unison I wouldn't have answered as I don't think there is one.
I took a quick look at the features page, and saw 'protection against media errors' and merging mentioned. There's much more, I get the impression it's the (command-line) infrastructure for a modern (graphical UI) archive util.