restore should be more intelligent?

Consider a large partition that is backed up as partition image by splitting it into 4GB chunks. So the backup files are from image.000 to image.02e

Let say the .028 volume is damaged. Instead of the restore failed by throwing out a file i/o error then stopped. Why can't the restore offer to restore the rest of the data?

From my understanding, an image contains all information of the said partition sector by sector in sequencial order. So it should be possible to skip the broken part by calculating the offset and continue the restore of the remaining volumes.

Chances are there are important data in the remaining volumes that are not damaged and are still usable. In my example, a client is forced to lose roughy 30GB of data because the restore stopped at .028 when it got an error. If the restore could continue, the client will only lose whatever data contained inside .028 volume which is roughly 4GB of data.

Now imagine same situation but this time .010 volume is damaged instead. The outcome is a disaster. Client using the backup because of a disaster and hope for the best but to find out the backup didn't help much. That's not a very nice restore mechanics, isn't it? Something to think about.