<$BlogRSDURL$>

What Do I Think?
Hello there. What do I think? Read on, reader.

Recent Posts

Best Posts

Links

Powered by Blogger
Visits:
Featured post: The Truth About The Half-Blood Prince

Friday, January 14, 2005
 
Book Recommenders
I'll use Amazon as an example in this post. On Amazon, there's the Editor's top picks and there's the Customer's top picks (bestsellers). These are good because they help us to decide which books are worth reading and which aren't. However, we often find that one person's favorite book is someone elses least favorite. Next, we have the categorized bestseller lists. The problem is that not every book in one genre is the same. Stephen King fantasy is worlds different than Terry Brooks. Plus, they put Sci-fi and Fantasy together, making it even harder. So, Amazon also has lists made by the customers themselves. If someone sees some books they like on the list, then they might trust the other books on the list to be good. However, this isn't gauranteed. Amazon also has the "Customers who bought this book also bought" thing. This is useful, but just like the former solution, it's not always what you want. Lastly, they record your purchasing and browsing history to shape the sidebar recommendations to you. The problem with this is that sometimes you'll get a book for school, or for a friend, or you'll glance at a book and Amazon automatically assumes that you like it. You have to go in there and specify the fact that you don't like it or don't care about it. Better yet, people change moods often. One week, they want mystery books. The next week sci-fi. So sometimes it recommends the wrong type of book.
The ultimate solution to this problem would be to have an entire site dedicated to book recommending where the user would create a profile and list his/her favorite books and categorize them how he/she wants. The results of his/her searches would be based off of other people's lists who have similar books in them. Also, the user can find books based off of not all of the books on his/her list, but only the ones that have the mood he/she's looking for at that particular moment. That way, the user can customize his/her recommendations how he/she wants without too much trouble.
I don't think anyone but I have thought of this solution quite yet. If no one else implements it, I guess I'll have to start it myself. And believe you me, I will.


Comments: Post a Comment
About This Blog
I Am Shmack
Call me Shmackshmael. Shmack for short. This is my blog. As you can probably guess, I am anonymous. No one knows my identity and no one ever will, though you may be familiar with my other blog, Shmack Reviews. If not, that is where I post my book reviews.
All I give are my thoughts and some small details about myself. If you're new here, please read "Welcome!" for more information on this blog.
Feel free to comment to your heart's delight. Your comments will keep this blog alive and inspire new ideas. Also, I will not give my e-mail address, so comment instead. I don't care if your comments have nothing to do with the corresponding post. If you become a daily reader, please comment and tell me just that. One-liners are fine. Thanks!

-Shmack

Copyright 2004 Shmack. All Rights Reserved.