Yes. It seems logical to tell authors to use useful numbering schemes rather then work around it.
Any person who sees Foo, version 1.3 and then Foo, version 1.3.2 would think that 'ah-hah' they must if released a bug fix version or something! Since, you know, 2 is a higher number then nothing which in human idioms is usually simultaneous with zero..
To me this seems very similar to trying to work around kernel bugs in userspace. Although in this case the bug is introduced into the metadata by upstream rather then program logic.