I see the appeal of the principle, but I think life of the author should not be a consideration in calculating copyright term, as it brings several problems:
1) In order to determine whether a work is under copyright, one needs to know not only about the publication history of the work itself (which is typically printed on copies of the work or accompanying material), but also whether and when the author died. For works by obscure authors, this can involve significant cost.
2) Works published late in an author's life or by an author who dies young receive a shorter term, even though there's no reason to think that they're less valuable.
3) It theoretically creates an incentive to kill the author to shorten the term. Although I suppose that this is quite unlikely to actually happen, why build such a perverse (in both the economic and moral senses of the word) incentive into the system if we don't need to?
Furthermore, it's not at all clear that term lengths on the order of a lifetime create any productive incentive compared to shorter term lengths. For the vast majority of works, revenue rapidly decays to a trickle within a few years of publication (for fast-paced markets like video games, this can happen in a few months). So starting from a term that's already decades long, an extra decade or two on a new work has a vanishingly small net present value [1].
[1] https://www.law.cornell.edu/supct/html/01-618.ZD1.html ; See Part II—C