Musicplasma.com

[Tom]

If interested in music, have a look at the Musicplasma (http://musicplasma.com/) site. I came across it on the Amazon.com Associates site, where it is given as an example of Amazon Web Services and give the below explanation. Whether or not this interests you, Musicplasma deserves a mention here as a specialist web-searching facility. David: its another example of web geography in action.


Web services is referred to as an API, or application program interface. An API is a set of building blocks made up of routines, protocols, and tools that govern how individuals use the service. It creates a consistent way for you to access the information you need from Amazon to integrate content onto your Web site or in an application. AWS can enhance your Web site with Amazon product information and make more money through the Amazon Associates Program.

Increase your referral fee earnings: Associates who use AWS enjoy higher click-through and conversion rates.
Scalable integration: AWS provides a scalable way for Associates to integrate richer Amazon product content into their Web site. Instead of manually creating product links one-by-one, Associates can easily access, display, and link to any of the million of items at Amazon.
Great product content: AWS exposes most of the rich product content that makes Amazon such a great place to shop, including real time pricing and availability, product images, product descriptions, customer reviews, sales rank, and more. The more product content Associates display, the more likely their visitors are to make a purchase.
Flexible merchandising: AWS gives Associates complete flexibility in merchanding products in a way that best suits the design and layout of their site.
Remote shopping cart: Customers can add items into their Amazon shopping cart without leaving the Associates site. This gives a professsional e-commerce feel to the site and a more convenient shopping experience for the visitor.

[David]

I loved playing with MusicPlasma. The slight problem is that it doesn't appear to have any more inteligence built into it than the Google "people who bought this also bought..." But what it does do is present this information as a wider landscape so that it can be viewed in a wider context. I guess that as the database of purchase links grow, the more accurate will be the landscape.

The resulting logic behind the landscape is rather muddy. People buy music perhaps because it sounds like other stuff they like, or because it falls within the same genre or because it is associated with a common band member or simply for no better reason than they heard it and liked it. And then there are those purchases made for other people which bear no direct link to the individuals personal taste.

And so, we have the strange situation where entering "David Sylvian" in the search box results in a landscape where The Clash, Elvis Costello and Radiohead all appear much closer to David Sylvian than do Japan. That is clearly nonsense and demonstrates the problem of relying on the raw Google data.

I would say that MusicPlasma is a great idea and could go places. Imagine this Google data combined with Rock Family Trees and timelines. Then you'd have something trully useful.