I’d love that in some fashion. At the very least, i’d love to generate that personal ranked list for you from your comparisons. But if we were to use it as an input mechanism, it’s hard for the grading system to handle it in it’s current form.
We’ve actually discussed before at one point moving away from Elo entirely and focusing on personal ranked lists like you’ve described. The task then would be to combine all those personal user rankings into one global aggregate ranking. However, my understanding is that it’s a relatively open research question in mathematics on how to combine partial ranked links into one aggregate ranking.
Another way would be to generate pairwise comparisons from those ranked list. However, Elo in its present form couldn’t handle those generated comparisons due to the limitation around a single user being limited to ~6 grades a book (as you point out). We could perhaps address that issue with a change to a static Elo system such as bradley-terry which I’ll eventually do… just would take a quite a bit of work and there are lots of high priority things now. However, the static elo system could probably handle the comparisons generated from this interface.
@seanblue does bring up some reasonable interface issues as well. I think i’d always allow you to just do the comparisons in pairs (its easy) but then generate the personal ranked list which you could tweak.
I think this is an interesting idea though, and if you wanted to make a product request, we could chat more there too. 
