I have been wanting to build this for 5+ years as well, exactly as described: find clothing-item overlaps between what I say fits and what others say fits -- th

I have been wanting to build this for 5+ years as well, exactly as described: find clothing-item overlaps between what I say fits and what others say fits -- the items we don't share can be thus recommended. Who cares about the actual numbers? If I love 32-32 in Levis STF 501s, and 33-32 in Banana Republic Gavins... and you are likewise 33-32 in BR Gavins, chances are great that 32-32 in Levis STF 501s will be a win. But major problems: clothing even within brand-lines changes regularly. Annual or semi-annual updates. Manufacturer changes. I have found size changes within colors even (BR boot cut chinos in khaki versus brown; major differences). And then identifying the clothes in the first place. We going to be scanning bar codes? Also getting support from the retailers -- both on sharing data and collecting it (they have much better chance of getting data from customers -- attach the web address to the clothes tag). But this will compel more online purchases and reduce return-churn, so they have an incentive to participate and monetize this.