[staticautomatic]

Sure, but A) As implemented in code, the path doesn’t necessarily have to be fully explicit in terms of tags. You could look for a child with text containing rather than a specific class or id, for example (or get fancier with semantic similarity on tags or text). B) There’s a trade off between speed and fragility that could make a difference if your tree is deep enough that traversing it iteratively is slow enough compared to a long xpath that it becomes a limiting factor. Granted you don’t typically encounter this in a standard issue html tree but the lxml docs, for example, correctly note that xpaths can be way faster when the nesting is super deep.