Today I visited Google to talk about web spam. I had never been on the Google campus so although I was busy I did make a point to note many details.

Just about every meeting room I saw was video enabled. The spam summit met in a medium-sized room with array microphones hanging from the ceiling tiles, three video cameras in the back of the room, and one video camera in the front of the room for crowd shots. Smaller meeting rooms had phones with eye-level cameras.

Lunch was really good. I had a chicken breast pesto panini with roasted peppers, fruit, and a real, fresh-made rice krispies treat. The kitchen was well stocked with all sorts of drinks. They had Smart Water, at least 5 types of root beer, bottled Frappuccinos, smoothies, and Coke in a bottle.

I took some shots of the Google master plan. There is a long set of whiteboards next to the entrance to one of the Google buildings. The master plan is like a wiki: there is an eraser and a set of pens at the end of the board for people to edit and contribute to the writing on the wall. I liked all the checkmarks next to the master plan accomplishments such as “Hire Vint Cerf.”

Thanks Google for hosting everyone for the day and providing geek sustenance of WiFi, power strips, food, and even clothes.

I have been reading through some of the posts about Google Blog Search and have some new thoughts on possible infrastructure although nothing has been officially stated by Google.

I’ve read about how fast Google’s results come back. I would hope so, their entire index covers only about 90 days.

Powered by Google Fusion?

We already know that Google Blog Search is indexing only feeds, and the index does appear to separate from the main Google index. We also know that Google’s feed search index only contains posts since June 2005. We also know that Google plans to add a form for inputting feeds in the future.

Pictured above is the a form field available on Google Fusion, Google’s personalized homepage and feed aggregator. This service launched in July, so data back to June is certainly a good possibility. Given all of the information we do know, it appears Google Blog Search is based on the same set of data used by Google’s feed reader.

Update: Google just posted a page with information about FeedFetcher, the feed retrieval robot for Google Fusion. FeedFetcher disobeys robots.txt and other things that are different than Google’s claims for it’s Blog Search product, so perhaps I am wrong.

Tags: googlefusion

Google Blog Search is now live. I spent some time clicking around and searching for terms and phrases where I expected a certain result. Some early thoughts:

• Query time is very fast, at around half a second for most queries. The database only goes back two months, so a quick response is to be expected.
• The related blogs search for “George Bush” suggests “George Bush doesn’t care about ugly hoes” which cracks me up.
• No advertisements. Is this another “mistake” like Google claimed for Google News? I don’t think so.
• Google exposes results as RSS and Atom feeds in two default flavors.

I think the relevance algorithm in Google Blog Search opens up a new testing ground for search engine marketers. While Google has been rumored to have a six month sandbox for new sites or domains, they cannot afford to be biased against newer blogs, especially with an index going back to June. Search engine marketers now have a way to create new sites and test their performance in Google Blog Search before they make it to the “big game” in Google’s main search results.

Google’s entry into the space will definitely increase the exposure of the entire industry. Many more people have heard of Froogle but have not heard of Shopping.com, NexTag, Shopzilla, or PriceGrabber. When Google introduces a product the world takes notice, even if the product stays at #10 for a few years as is the case with Froogle. In the shopping comparison industry everyone continued to have a nice business even after big search companies entered. Only time will tell how Google performs and evolves. They may never remove the beta designation, but that would just be typical Google.

Google is launching a feed-specific search interface tonight for both all feed producers as well as a search interface specific to Blogger.

It’s important to note that from what I have read on Search Engine Watch and others Google is specifically restricting its search to feeds, and not using the HTML of the blog. Why? Googlebot is designed to swallow a page whole and not break the page up into individual entries or items. Feeds come prepackaged as individual items or entries allowing for easy digestion by parsers and indexers. Google would need to overhaul its indexer or design a new and separate indexer specific to blog posts if it would like to include more post content than it is currently pulling down from a page’s link alternate declared feed (this is based on a conversation I had with Google engineers in February about the indexer, I won’t blog the details, and things may have changed). Technorati indexes a blog’s HTML assisted by the declared RSS and Atom feed, so I am admittedly a bit biased.

I think feed search is just another type of search restricted to a group of MIME types, as previously stated, and as the types of content made available via a feed continue to grow I think a feed-specific search tool will become much more than just blog search.

More on the actual Google products as they become available.

Update: Google Blog Search Help is now online.