Learnings about Search Engines

From Computers in Libraries 2006

 

AskJeeves has become Ask.com, but more than the name has changed. Four of the five major reviewers of search engines say that Ask is now the leading search engine in terms of best meeting users’ needs. (e.g., see Wall Street Journal http://ptech.wsj.com/archive/ptech-20060330.html).

 

All of the search engines are seeking to become more than just search tools, but information portals and even more. For example, see live.com, yahoo.com, or personalize your google home page (www.google.com/ig).

 

While there are thousands of search engines, been acquired, merged, or otherwise dropped out. There are now four major search engines—

1.     Yahoo! is still #1 in page visits (including all offerings). Their search is on a par with the other three majors. Their beta offers an autocomplete instant results bubble through ajax. My Web! is a social bookmarking database that tracks their users’ recommendations as bookmarks. It also allows you to save pages to their server (a personal cache). Results from the same domain (e.g., familysearch.org) are limited to only one hit. To get around this, add &dups=1 to the end of the search results URL and press enter. For advanced search help, go to help.yahoo.com, then click on “All tips” to see the useful stuff.

2.     Google has a unique method of relevance ranking that shot it to the forefront as a search engine. They rank pages according to how often it is linked to by other pages within the search results. While usually positive, there are drawbacks. For example, the actual search term may not appear on the resulting page—only in the links to it (to get around this, use intext:). They have 52 servers and each is a little different. Just reload and you may get different results—see http://www.webrankinfo.com/english/tools/google-data-centers.php).  To get around the 1 hit per domain limit, add &filter=0 to the end of the search results URL and press enter. Google cache is wonderful (see what the page looked like when they indexed it). If Google cache page says “supplemental result,” it is a really old one. Google is constantly experimenting with new features or interfaces. You may see something and go to show it to someone else, but you may never find the same feature again. Use standard syntax to use google search as a calculator--e.g., (127.5*13)/1.5. For more information on google advanced search, see www.cwire.org/data-mining-using-google/ or google CheatSheet at www.google.com/help/cheatsheet.html. Check out Google Suggest at www.google.com/webhp?complete=1&hl=en. 

3.     Ask.com has the most successful natural language search, presentation of related data, and suggested links that are relevant to your search. Maps have the best zoom, and can overlay street maps on satellite imagery. Check out the features of the search tools box to the right. Use the advanced search, which is unfortunately hidden on the next screen of search tools, or advanced help at  about.ask.com/en/docs/about/site_features.shtml .

4.     MSN search (becoming live.com) is slightly smaller than the other three. It has two useful unique features – search results link to Encarta. When clicking on an Encarta link from your search results, you automatically get 2 hours of free access to Encarta (normally a fee-based encyclopedia). Also, their maps are the best. For example, you get different directions depending on whether you are walking or driving (you CAN walk the wrong way down a one way street).

 

Sometimes, you would like to search a specific site and are having trouble using the search box on their site (or they don’t have one). Just insert site:edu at the beginning of the search box, then your search term.

 

While each of the search engines seeks to have up-to-the date results, sometimes the links are old. With Google, click on “cached” to see just how old the results are.

 

Possible future enhancements:

·        Take a look at Exalead. It has some great features that undoubtedly the majors will copy to prevent Exalead from stealing more than the 6% market share it has grabbed in the past 12 months.

·        Gigablast allows you to create custom searches. Enter the URLs of up to 200 web sites, and Gigablast will create an index that will allow you to restrict your search to those 200 domains.

·        A9 is a scaled-down version of Google for Amazon. It searches for book content. You can search the content of many current books. It has superior results zoom features. Also, if you find a book using A9, you get an automatic discount if you then buy it through Amazon.

·        Try Soogle (incorporates advanced search into basic search).

 

Alan Mann, April 2006