I wanted to be able to link to any page of a PDF in my Markdown notes, and I knew that I could do it by appending #35, for example, to the end of a file:// URL in my browser. But I use Skim, and reading PDFs in a browser is no good. Further, the URLs end up being super long with ugly things like %20 for spaces and this seemed unnecessary because all my PDFs are carefully organized with BibDesk. So I set up a little AppleScript that would parse a custom URL starting with sk:// for a BibDesk citekey and a page index, then figure out which PDF I want from BibDesk and open that PDF in Skim to the right page index.

To get this to work I need to save this using Apple’s Script Editor as an independent App called Skimmer. Then in the info.plist for this App, I put this:

This makes OS X recognize that URLs starting with sk:// are Skimmer URLs.

So when I put a citation reference like [@mccabe94 [page 41](sk://mccabe94#57)] in my markdown source, I end up with a footnote that looks like this:¹. Clicking on the link in this citation takes me to page 41 of Plato’s Individuals. The script first pulls out the citekey mccabe94 and finds the entry with this key in Bibdesk. Then it opens the first linked PDF of this entry in Skim and tells Skim to go to the page index after the #. This has the advantage that I can decide to reorganize how I store PDFs in the future and all my links will be unchanged.

Notice that there is a difference between a PDF page’s “label” and its “index” the index is an integer with the first page of the PDF counting as 1. The page’s label is a string and is often used to label the cover of a book or pages that use roman numerals. Hence in McCabe’s book, the page with the label “xii” has the index 14, while the page with the label “4” has the index 20. URLs to specific pages always use the index because these are unambiguous, but when people refer to the “page number” in a book, they usually mean the page label.