Yes, I should have mentioned that only articles that are shared (public) and that have extended description text are processed by my script. That way I don't wind up with any articles that have no body content and just a title and tags.
I see. I was wondering if you were filtering your bookmarks by specific tags, but excluding description-less bookmarks also sounds like a good idea.
BTW, your scroll.js has a commented-out section where you call 'pinboard.php', so I suppose you're using PHP to process your bookmarks. As the author of the PHP Pinboard API Client (github.com/kijin/pinboard-api), I'm curious whether you're using it, and if so, whether you have any suggestions for improving it.
I'm not, but I'll definitely check it out. Since this application is so simple, it only requires one curl request and I just use the http authentication method. Working on this small project (and the comments here on HN) have definitely opened my eyes to greater possibilities though, so I'm sure I'll have need for a php client in the future.
