Laurabelle’s Blog

I've been playing around with What Should I Read Next? which is interesting but not fully functional. I'm not just referring to the fact that the site doesn't yet have a critical mass of data for recommendations; that's understandable. What's annoying is that the interface has some bad design choices that I don't understand.

The site's killer app is that users can create a profile of books they have read and liked; this mass of book associations gets chewed up and spat back out as recommendations. So far the recommendations aren't very good, but that has a likelihood of improving with more users and more data. So far so good.

I have a few criticisms, which I will enumerate in list form.

1. There's no password authentication. If you know the email address I used to register, you can access my book list. This is a horrible no-no. Now, it just so happens that I don't care if the whole world knows I've read Agatha Christie's And Then There Were None, but really, is a password too much to ask?
2. There's absolutely no sorting, grouping, or pagination in the list of books in one's profile. For bibliophiles like me, this becomes a problem very fast.
3. Adding books to one's list of read-and-liked books is very annoying. The little author/title form is way at the bottom of the book list, and if the author and title you give don't exactly match an item, you have to click a link on the next page to confirm that what the site thinks you're looking for actually is right. However, you only get one search result to choose from and no pointers to why that was the one chosen. If it wasn't what you were looking for, tough luck.
4. The only information available about a certain book is author and title; there's not even any information about format etc.
5. When displaying recommendations, the site doesn't display the book(s) one used to search. This is annoying for someone as occasionally scatterbrained as I. It really wouldn't be hard to display the search query, right?
6. One can only mark books that one has read and liked, not books that one has disliked. This is, I think, an important omission.
7. The unique identifier for each book seems to be ISBN. This isn't a great idea because ISBNs are not actually unique. Publishers will sometimes re-use them and cause merry hell for librarians and other people who have to use them.
8. There's no attempt to identify different editions (or manifestations or instantiations) of the same work, for example the compact disk audio book or the Spanish translation of Harry Potter and the Order of the Phoenix. Or the two titles of Ten Little Indians and And Then There Were None, which are exactly the same book. This is rather hard to institute without good (manual) cataloging, but the idea doesn't even seem to have occurred to the people who designed the site. For the purposes of this site, pretty much anything falling under the same work should be considered identical.

Those are the criticisms I have right now. The last three are relatively minor; I could be pretty happy with the site if those were the only issues. However, as it stands, the site just isn't usable. It's a royal pain to add books; why can't I just search for Terry Pratchett (almost all of whose books I have read, dozens of them) and add everything at once? Instead, I have to search individually for every single item. Dumb. If the site stays this hard to use, they'll never get enough data to build worthwhile recommendations.