Now here’s a search engine destined for controversy. Pixsy.com, a “visual search engine” that scours the web for images and videos on RSS feeds and makes them searchable, launched yesterday – for the second time, and it may not be ready still.
Pixsy‘s maiden voyage was last July, but due to limited partnerships and a reliance on XML, it couldn’t deliver on its promise. Founded by four pedigree engineers hailing from Microsoft, Sony, InfoSpace, and Corbis, the AJAX-based website has been reconfigured to move beyond XML, and has procured partnerships with The New York Times, BBC, YouTube, and Business Week among others.
The idea is to “explore the web visually” as Pixsy’s search engine claims to combine freshness and relevancy – and only returns visual results with a listed source.
“Anywhere there’s an RSS feed, we consume it, extract an image…and make it searchable,” said Chase Norlin, founder of the company.
Image search has been on a serious upswing over the past year as vertical search services become popular among more sophisticated surfers. Still, Google has had issues with publishers over indexing copyrighted material. Pixsy has covered that base, in large part, due to its partnerships. But, it remains to be seen how publishers outside that partnership will react.
The site is supported by contextual pay-per-click advertising and its partnerships, and has future plans for licensing to other companies. The partnerships it already has, like with The New York Times and the BBC, provide a very interesting photojournalism experience with the latest photos from those publications and the ability to click and find the articles associated with them. That’s one aspect that makes this service stand out among other image and video search engines.
However, where it already has trouble is in relevancy. A search for Paris Hilton, for example, brought up few (or none) images of her on the first page. Instead, they were photos of other people, or a little white dog (a Bichon Frise?) written about in a Spanish blog post that mentions Paris Hilton. (Curiously, a subsequent search was without this problem).
Pixsy also seems to have an issue with Firefox. When tabbing two separate photo searches, leaving the first to go to the second, and then returning to the first (are you lost?), the first tab takes on the subject of the second one and you’ve lost the original. For example: leave Paris Hilton tab for Jessica Simpson tab, return to Paris Hilton tab and click the next page and its now Jessica Simpson.
Finding images seems very dependent upon what is current in the feeds that mention the keywords, and as someone else pointed out, searching for adult film stars or nude models may bring back very irrelevant photographs of flowers – along side an “error getting Looksmart ads.”
These are some issues the crew at Pixsy will have to work out or users will find little use for it. They have some big name partnerships, but aren’t delivering the goods 100% of the time. And it is unclear how Pixsy plans to avoid RSS spam, if a little keyword trickery can fool it at times.
Maybe some more tweaking will precede a third launch, and that will be the charm for Pixsy.com.
Tag: | document.write(“Email murdok here.”)
Drag this 
Add to document.write(“Del.icio.us”) DiggThis Yahoo My Web
