I'm listening eagerly but I don't know the answer.
Some questions:
Does the pollen affect fruit shape or other characteristics at all?
Does one grain of pollen produce every seed in one fruit, or is the rule:
"one-grain-of-pollen = one seed"?
>> It actually seems to me that the seed from all the tomatoes on one plant should have the same genetic makeup,
Well, say there is ZERO cross-pollination, which is possible, since tomatoes are mostly self-pollinating. Both parents would be the same plant ('Pruden's Purple').
However, recombination means that every seed will have SOME variation, since the 'Pruden's Purple' variety is not 100% homozygous in every trait. They would be as similar as rather inbred brothers and sisters, but not identical twins.
Since 'Pruden's Purple' is an heirloom with fairly stable genetics, it is fairly uniform genetically (highly inbred, mostly homozygous for most important traits). Hence the "brothers and sisters" analogy is weak. Compared to people or animals, OP plant varieties are VERY inbred, so 95% of the "brothers and sisters" come out very similar.
I think you can select for one particular fruit shape IF there are any plants that have mostly one shape. But what if the shape is very dependent on some non-genetic factor like weather, water or pure chance?