Cool! I like that link, because it confirms what my mind had worked out about a nifty H-frame matching circuit I saw once.
That design dispensed with the extra 1/2L section of coax, and just connected the two 1/4L impedance match sections directly to the center feed lines of the H-array, but 1/2L apart, getting the same end result for free.
The scary part is, I think I fully understood it.
So now, for my own 50ohm quad-dipole H-array, I just need to fabricate a pair of 1/4L matching sections with a 61ohm impedance (to connect the 50ohm H-array to 75ohm RG6).