Category Archives: Search Engines

Search Engines

Dogpile Missing Pieces

Dogpile Missing Pieces

A neat little tool from meta search tool Dogpile that compares the top 10 results from Google, Yahoo and Ask Jeeves. The graphic shows how many results are in only 1, 2 or in all 3 search engines for a particular search and includes sponsored links from the top of the pages.

Dogpile have also produced a white paper (PDF format) on the subject entitled Missing Pieces: A Study of First Page Web Search Engine Results Overlap. The findings are no great surprise to many of us but it is always useful to have our experiences confirmed by experimental data. I find that the degree of overlap varies with the search: sometimes it can be as low as 2-3% in the first 100 results whilst on other occasions it can be as high as 90%. If you want to compare a wider range of search tools and their top 100 results try Thumbshots Ranking.

Google, Search Engines, Yahoo

YaGoohoo!gle

YaGoohoo!gle

An interesting tool that runs your search on both Google and Yahoo at the same time. It displays your results in two fames alongside one another in your browser so you can view both sets at the same time. Actually – its better to look at them one at a time; I started to feel seasick trying to compare the results in both sets. If you just want to compare coverage and results between Yahoo and Google for a particular search strategy then use Thumbshots Ranking.

Search Engines

Gigablast Related Pages

Gigablast Related Pages

This is not like the Google related or similar pages where you start with a single relevant page and ask Google to find other pages similar in type and content. The Gigablast offering works at the search level: you type in your search strategy as usual and near the top of the results list there is a “related” pages section displaying a few extra pages with a more link to other “related” pages.

I found that they do not always appear, especially when you type in a complex or more detailed strategy. Also, I am a bit suspicious as to how these so called related pages are selected. Gigablast gives an example of searching on Colorado activities:

“You will see many webpages which are contextually related to the original query terms, but have no obvious direct connection to them. Many show the word Colorado, but not activities or activity, yet the pages all seem to fit well into the descriptive two-word query Colorado activities.”

And then:

“Upon searching for the original query terms within several of these pages, one is left wondering how the Gigablast software code is able to distinguish them as being relevant in the first place. The answers? …proprietary, of course.”

After running a few of my standard test searches, I have the impression that a lot of these pages could be paid-for placements and not many were relevant to my searches. But perhaps I am being uncharitable. More useful to me is the Giga Bits section, which shows alternative search strategies and they are relevant.

Google, Search Engines

BananaSlug

BananaSlug

Another gem from Phil Bradley’s blog (I don’t know where Phil finds these sites!). You put in your search strategy and then BananaSlug adds a random search term. Alternatively you can select a category for your random word – e.g. animals, great ideas, random number, themes from Shakespeare.

The idea behind this site is to promote serendipitous surfing. By adding a random term, which may or may not be relevant, you pull up pages that are buried way down in the results list and which you would probably never see. Sometimes it works, sometimes it doesn’t.

BananaSlug uses the Google APIs and is limited to 1000 queries a day. If the site is past the limit, you are diverted to Google with your search term and random word.

Search Engines

Brainboost Answer Engine

Brainboost Answer Engine

I groaned when I saw that Brainboost “uses Machine Learning and Natural Language Processing techniques” to answer questions written in plain English. But I found this “answer engine” to be very good. It successfully answered “which US presidents were assassinated”, “which UK prime ministers were assassinated”, and “who directed the Duck Hunt cartoon”.

It was not so clever on “who is Karen Blakeman”. Brainboost came up with some 3 year old information on a course I was running at the time in London, and ” Karen Blakeman is chairing the Concessions Committee, and will be needing LOTS of help here”. You bet I will! I know nothing about the Concesssions Committee – obviously another Karen Blakeman. The regular search results taken from a range of search tools fared better on the last one.

Worth a look and an interesting alternative answer/reference tool to answers.com.

Search Engines

monzy.org Unsafe Search

monzy.org Unsafe Search:

I was alerted to this by fellow Internet Consultant Phil Bradley via his blog.

Unsafe search works by performing two Google queries, a normal query and a query with SafeSearch enabled. The set of ‘safe’ results is then subtracted from the set of normal results to yield only ‘unsafe’ entries. The idea behind Google’s SafeSearch is to eliminate p**n and other “undesirable” sites.

What is interesting about Unsafe Search is discovering what Google considers to be “unsafe”. The first test search I ran was on air quality. It yielded only one unsafe result in the top 30: the Idaho Department of Environmental Quality! A search on chocolate consumption came up with 3 unsafe results (one on chocolate is good for pregnant women, one on the potential health risks and a link to a Nestle forum). A search on gin vodka came up with no unsafe results in the top 30. Great! Now where did I put that cocktail recipe book?