The current way that the validator processes block nodes is to start from the last sibling and progress backwards to the first. It was set up this way originally because it made it slightly easier to deal with repeatable node annotations. It turns out that it increases the difficulty of interpreting validation errors though since it is not intuitive to start at the end and work backwards. If the order could be reversed, it would increase the sensibility of the reported errors and would make the validation routine consistent between block and inline node types.
The current way that the validator processes block nodes is to start from the last sibling and progress backwards to the first. It was set up this way originally because it made it slightly easier to deal with repeatable node annotations. It turns out that it increases the difficulty of interpreting validation errors though since it is not intuitive to start at the end and work backwards. If the order could be reversed, it would increase the sensibility of the reported errors and would make the validation routine consistent between block and inline node types.