Infoseek Storms GeoCities Pages

In an attempt to maintain the accuracy of its index, the search engine severs a domain to remove the abusers. No, it's not nice to fool Mother Crawler.

When attempting to submit URLs to the index of search engine Infoseek, members of free homepage service GeoCities have found themselves rebuffed of late. The reason: too many members trying to fool the engine into giving their pages priority.

In an attempt to float a page to the top of search results, authors can employ tricks to make the page appear disproportionately relevant. Repetition of certain keywords, for example, ("sex" being a perennial favorite) and loading a page with hidden, but frequently searched, text are some of the tricks used to increase a page's visibility. Search engines can be trained to retard this type of "page-loading," but the wily page will find its way.

With more than 500,000 members taking advantage of GeoCities' advertiser-supported page-hosting, Infoseek made the service a target simply because doing so cast a big net - lots of users, lots of abusers.

Infoseek spokesman Steve Grady said the service is not universally removing GeoCities sites from its index. "It's not necessarily that GeoCities was singled out," he said. "We've done it with other ISPs." He wouldn't name providers targeted in the past, but said Infoseek had seen a large number of "spamming" pages coming from geocities.com. "We'd exclude any pages that try to affect search results," Grady said. "They're attempting to subvert the index."

Furthermore, Grady said, GeoCities' members were never prevented from proffering their pages through the service's email-based submission process. It is only the site's automatic URL-submission procedure that has been closed off to GeoCities domains, he said.

GeoCities representatives were not available for comment Monday, but Grady said the two sides "are working together to identify how it's happening and ways to control it." When Wired News submitted a query to Infoseek support, the reply said the company was "currently working with GeoCities ... this will only be a temporary block."

Grady said many pages are removed as part of Infoseek's routine and ongoing effort to weed out loaded pages, but not GeoCities' exclusively. "We need to do everything we can to ensure that when our users conduct a search that they get back relevant responses to the topic," he said. "So we've taken a very aggressive approach to doing that."

Authors whose pages have been removed can have their pages re-indexed "if they agree to abide by the guidelines," Grady said. They must send Infoseek email stating that they'll do so.