[alialkhatib]

I get the impression that if this wasn't calling on a third-party API (ie if you were indexing the product listings yourself) you'd be able to give more "instant" results, but this is a cool take on search on Amazon (although obviously inspired by Google and other instant search features). I also imagine Amazon would blacklist you for harvesting their product listing, although since you're still giving them referrals maybe they wouldn't. I guess if you called it "aggressive caching" you might get away with it, heh.

While you're re-thinking product search, why not consider higher-resolution/larger thumbnails, a few more products per "page", etc...? I'm not advocating a Pinterest clone or anything like that, but the initial load (10 items in 2 rows) looks a little... dated and spartan.

As said before, cool short-term project. Would be interesting to see a few more features built into this, especially if it made your search "competitive" with Amazon's implementation.

[lugg]

Amazon would not ban you for this. They encourage it. They have throttling on though.

[fjw]

I think it would be really interesting to download the entire listing and watch how volatile everything is, especially the item prices. Searching would definitely be much faster as well.

Thanks for checking it out and thanks for the suggestions!

[alialkhatib]

That's a really interesting idea; tracking prices and potentially giving a Buy/Wait indicator would be really cool.

CamelCamelCamel[0] actually does some price tracking, but I'm not very familiar with what they do with it (with regard to giving advice on whether to buy or not). You might be able to grab some information from them on products via an API or scraper. I'm not seeing an API, nor do I see anything about using their data for other stuff, but I suppose it goes without saying that a really aggressive scraper wouldn't be welcome.

0: http://camelcamelcamel.com/

[xur17]

I've used CamelCamelCamel a few times - they let you choose a price, and they'll send you a notification when the price drops below that. Something that can determine when the price is at its lowest based on previous trends would be pretty neat.