Nicole B. Cignoli - Capstone Portfolio

Web Exercise – Unit 5

July 7, 2007

 

To complete my search about meteorites, I chose the following three web search engines: Yahoo!, Google, SciSeek, and metasearch engine, Dogpile. As Google and Yahoo! are rated the top two most accessed web search engines by www.searchenginewatch.com and Dogpile as the best metasearch engine, I felt I would find greater information and links to understand meteorites’ composition, value, and to locate historic and recent impact sites as well as where to sell. Sciseek is a focused web directory created by humans and reported to be the best guide in the field of science and nature, which I thought would help define composition and historic impact sites.

 

Search Engine Comparisons

 

Yahoo!, Google, and Dogpile sites have advance search links easily located on their home pages. Yahoo!’s and Google’s advanced search tools offer form fields to search for “all of these words” “any of these words”, and “none of these words” as well as time period, domain, and web sites. Yahoo has a tip for each section of the Advanced Search field to assist the user to obtain a more precise search. At the bottom of the page, Yahoo!’s Search Tips explains a Boolean search operator with additional tips and suggestions on how to conduct a better search on their engine. Sciseek’s advanced search link is not as easily accessed from the home page not does it provide clear usage instructions. However it offers the same functions.

 

All engines process multiple word searches using AND so adding more terms narrows the results with the exception of very common words such as "it" and "how”. All use double quotation marks for exact phrase search and can create phrases using punctuation or special characters such as dashes, underscore lines, commas, slashes, or dots. Yahoo! supports full Boolean and nested searching with the operators AND, OR, NOT, AND NOT or NOT as well as math language (+ to include, - to exclude, etc.). Google and Sciseek translate the abovementioned operators into math language, which is their operator. Boolean operators must be in upper case. Unlike Google, Yahoo!, Sciseek, and Dogpile are not case sensitive so searches can be typed in lower or upper case. When you use upper case text in Google, the search service finds only upper case. Google, Sciseek, and Dogpile use and (*) at the end of a word to search for the word with multiple endings. Yahoo! uses a stop word “a of” at the end of the word in question.

 

Yahoo! and Google include key words and cached copies of pages to quickly link the user to the document or site. Dogpile searches various web search engines and provides the link to the hit with a reference at the end of each hit of all the engines that contained the same link. All search engines remember key word searches. Yahoo! lacks some advanced search features such as truncation, only indexing the first 500 KB of a Web page (still more than Google's 101KB), linking searches requires the inclusion of the http://, file type searches use originurlextension: rather than file type, and including some pay for inclusion sites. Sciseek’s strength is also its weakness. As a science search engine, I was unable to locate peer-to-peer information or analysis. Dogpile included advertisements but placing -ads at the beginning of the key word search moved these links lower in the search results.

 

Search Results and Analysis

 

	Yahoo!	Google	SciSeek	Dogpile
meteorite	2,740,000	4,450,000	30	98
meteorite identification	140,000	1,560,000	40	60
meteorite showers	73,700	1,470,000	30	20
meteorite hunting	188,000	203,000	40	75
meteorite hunters	79,800	218,000	30	87
rocks, gems, minerals>meteorites	47,000	95,100	20	88
rocks gems mineral meteorites	47,200	95,100	40	84
rocks gems minerals meteorite + hunting	1,820	43,900	29	78
rocks gems minerals meteorite + hunting strewn fields	24	717	2	58
meteorite AND (contraband AND illegal)	239	783	13	78
meteorite AND (contraband OR illegal)	134,000	206,000	13	76
meteorite recovery	147,000	255,000	38	84
meteorite (recovery + observed)	28,200	73,800	40	68
meteorite (recovery + find)	80,600	140,000	30	74
meteorite recovery permits	1,190	51,900	30	74
"meteorite recovery" litigation	28	9	30	21
Sell contraband meteorite any domain in the last year – Advanced Search	194	714	18	49
contraband meteorite treaty	61	497	14	65

(Back to the top)

 

Based on previous searching experience with Yahoo! and Google, I expected the preliminary search in these engines for meteorite would yield a large number of hits. I was pleasantly surprised when Sciseek had 30 initial hits for meteorite and Dogpile had 98 having never used either tool. Yahoo! suggested additional searches in meteorite identification, meteor showers, meteor hunting, meteors, meteors and impacts and quick links to fall phenomena, meteorite types, and meteor recovery. Some of Yahoo!’s hits offered category information of Rocks, Gems, Minerals>Meteorites. I added these suggestions to the end of my preplanned search terms. Dogpile offered a call out box titled, “Are you looking for?” which suggested terms and categories. Google and SciSeek did not offer categories or terms.

 

Google had the largest number of hits in every search category. I assumed the initial search results in Yahoo! and Google would produce many irrelevant sites. However, both engines’ first ten of the first thirty hits in the initial meteorite search provided a blend of research sources in definition, impact information, and educational and buying and selling sites. Google and Dogpile’s first thirty sites in the initial search yielded 80% precision compared to Yahoo!’s 60% and Sciseek’s 77%. All search engines had the same two false drops. The first looked promising as Meteorite, a quarterly meteorite magazine, listed at www.meteor.co.nz. However it was actually a financial institution located in New Zealand. The second was more easily discernable as a false drop under the heading Meteorite providing www.meteroritemusic.com, a website for a Birmingham based funk band.

 

Yahoo!’s, Google’s, and Dogpile’s (without –ads) first hit for meteorite was www.wikipedia.com’s definition of a meteorite while Sciseek listed it 28^th. Yahoo!’s second hit was www.nineplanets.org a site that provides complete charts of meteors and links to other sites for sources, selling and buying. The third hit was www.meteoriteexchange.com (a site to buy and sell meteorites), and the fourth was www.encyclopediabritannic.com. Google’s next three top hits were www.seds.lpl.arizona.edu/nineplanets, www.meteorite.com, and http://eps.www.unm.edu/iom/howto.htm. Dogpile’s next three appeared under the heading Meteorites for sale as follows: 2) a site sponsored by www.arizonaskiesmeteorites.com, 3) www.meteorite.com, and 4) www.meteorite.Tv/ a private organization which supplies meteorites for science education and collectors. Sciseek’s first hit was www.meteorite-times.com, a magazine published by The Meteorite Exchange. Sciseek offered many valid links to sources, historic impacts, American Meteorite Society, Jet Propulsion Laboratory, and Arizona State University. Although it was purported to be science related, it also had links to buying and selling websites such as www.alaska.net and www.meteoriteexchange,com, as well as art galleries, and modern-day hunters.

 

The most interesting site was www.saharamet.com, which explained that in Northwest Africa, meteorites are collected to fund terrorist cells in Morocco and Algeria.  I attempted pursued this idea by searching Contraband AND illegal which located illegal links in all subjects, not just meteorites. “Meteorite recovery” litigation found 9 links in Google referencing litigation about Ford trucks. In all attempts (and more not listed), I was unable to locate a valid hit from United Nations or International Law. (Back to the top)

 

Evaluation  

 

In order to search and compare sites, I created a pre-search list of terms: meteorite, meteorite hunting, hunters, profit, clubs, trip, historic. I changed my terms based on Yahoo!’s suggested terms and categories.  I began my search on Yahoo! through all terms systematically cataloguing  each search’s first 30 hits.  Then I crosschecked those results with Google, Sciseek and Dogpile in successive order. As illustrated in the search analysis table, the narrower my keyword search, the fewer the hits. Yahoo! and Dogpile reduced dramatically to equal the numbers and sites Sciseek originally located. However, the search positioned the sites in different order or located false drops. Switching word order didn’t locate new sites, only dropped hits.

 

The use of the stop word functions sped up typing and enabled the engine to search more possibilities. Using Booleans operators moved hits to different positions within the overall search and did not always produce new results. I believe I narrowed the search too much even though Google couldn’t locate less than under 100 valid hits. The two sites on SciSeek, www.rockhoundstation1.com (a global rock hound community website) and www.desertusa.com (a site for the southwest with links on how to identify meteorites, books, tours, etc. to help the enthusiast locate meteorites) are good results. Alternatively, the first five of twenty-four hits on Yahoo! were valid, although repetitive of previous finds. Then the search produced false drops for books on Turkish history, Mi Casa Su Casa Bed & Breakfast, blogs, and an Internet game.

 

Overall, I was satisfied with the web sites located. Sciseek offered the greatest number of relevant sites in order of importance. I would use Sciseek as the foundation of my research and add the sites found on the other engines to complete the analysis. My attempt to reduce the number of hits prolonged my research and clouded my analysis that the most varied selection of sites to meet my goals were found in the first thirty or fifty sites during the initial meteorite search on all engines. With more experience, I will be able to search more successfully and precisely using more appropriate search engines and with a better command of Boolean operators.   (Back to the top)