I created a drum machine to interface a rhythm database.
There are as many pages as stored rhythms, each with a friendly URL containing genre and rhythm name in it. Them all pages look almost identical, the only difference is on the rhythm details: genre, name, author, and accents pattern.
Most of the text of each page is duplicated, specifically I have a few modals for: the sounds collection, the rhythm list, the inline help and the about section. The rhythm list, in turn, is paginated. These are included on every page.
Considering all that explained above, I decided to set the root path as the canonical URL for all the other pages. I.e. only one title, no different titles for every rhythm's page. Am I right doing so? I think so.
The thing is, now I would like to include microdata with the WebApplication schema (http://www.schema.org/WebApplication) for the drum machine, maybe also with the MusicComposition schema for the rhythms, but only some rhythms (e.g. 15) are included in the root page when it first loads. The other rhythms would remain unknown because of the pagination (items are loaded with AJAX). How could I get all the rhythms catalogued?
Lastly, I have a doubt relating all this (in case I create a new page listing the rhythms): I have read that for paginated content, e.g. a post consisting of 3 parts, I should create a page containing all content together and set that page as canonical for the others. Now, canonical is a way to say "preferred" for indexing purposes but, what if I prefer the parted pages to be indexed, i.e. I don't want the aggregated page to be accessed, but any of the pieces.