That sounds like a great idea.
Instead of "legacy macros," I would use the term "unhygienic macros." That already implies enough of a value judgment, and "legacy" implies that we've left them behind. Frankly, hygienic macro systems still have enough problems with non-trivial macros that I sometimes long for the simpler days of generate-uninterned-symbol.
I would especially like to see examples of real-world macros — not the trivial examples that are so common in academic writing and Scheme implementations' documentation, but ones that push the limits of what can be done, but whose complexity is justified by the fact that they bring simplicity to the users of the macro, e.g. by defining domain-specific languages.
Thanks for doing this.