72-et 11-limit blocks

I calculated a reduced basis for the 72-et in the 11-limit, obtaining
<225/224, 441/440, 540/539, 9801/9800>. I then added 81/80 as a step
vector, and got a notation [h72, -h14, -h50, -h12, -h31], which tells
us we can produce a Fokker block by taking

(81/80)^n * (225/224)^round(-14n/72) * (441/440)^round(-50n/72) *
(540/439)^round(-12n/72) * (9801/9800)^round(-31n/72)

as n goes from -36 to 35. I did this, reduced the results to an
octave from 1 to 2 instead of from 1/sqrt(2) to sqrt(2), and obtained:

[1, 99/98, 45/44, 77/75, 28/27, 21/20, 35/33, 15/14, 264/245, 49/45,
11/10, 10/9, 9/8, 112/99, 8/7, 231/200, 7/6, 33/28, 32/27, 6/5,
40/33, 11/9, 99/80, 5/4, 44/35, 14/11, 9/7, 35/27, 21/16, 33/25, 4/3,
66/49, 15/11, 11/8, 112/81, 7/5, 99/70, 10/7, 81/56, 16/11, 22/15,
49/33, 3/2, 50/33, 32/21, 54/35, 14/9, 11/7, 35/22, 8/5, 160/99,
18/11, 33/20, 5/3, 27/16, 56/33, 12/7, 400/231, 7/4, 99/56, 16/9,
9/5, 20/11, 90/49, 245/132, 28/15, 66/35, 40/21, 27/14, 150/77,
88/45, 196/99]

Since 72 is even, some choice needs to be made on the boundry cases,
but aside from that this should not depend on the choice of 81/80. As
a check, and to see what I would get, I redid the calculation using
126/125, which gives the notation [h72, h58, -h50, -h12, -h31]. I
then got

[1, 99/98, 45/44, 77/75, 28/27, 21/20, 35/33, 15/14, 264/245, 49/45,
11/10, 10/9, 9/8, 112/99, 8/7, 231/200, 7/6, 33/28, 25/21, 6/5,
40/33, 11/9, 99/80, 5/4, 44/35, 14/11, 9/7, 35/27, 21/16, 33/25, 4/3,
66/49, 15/11, 11/8, 112/81, 7/5, 99/70, 10/7, 81/56, 16/11, 22/15,
49/33, 3/2, 50/33, 32/21, 54/35, 14/9, 11/7, 35/22, 8/5, 160/99,
18/11, 33/20, 5/3, 42/25, 56/33, 12/7, 400/231, 7/4, 99/56, 16/9,
9/5, 20/11, 90/49, 245/132, 28/15, 66/35, 40/21, 27/14, 150/77,
88/45, 196/99]

As expected this is very nearly the same. I was interested to see I
got 28/15 instead of 15/8--alas for the diatonic JI!

--- In tuning@y..., genewardsmith@j... wrote:
>
> Since 72 is even, some choice needs to be made on the boundry
cases,

I advocate centering on the 1/1-3/2 fifth in even cases. I couldn't
figure out how to make my PB program do that, but I didn't try very
hard.