Search Engines and the Invisible Web Lecture

advertisement
INFO 522
Lecture 9
Missy Harvey
Winter 2011
Search Engines and the Invisible Web
Many libraries have created tutorials to help people learn how to search on the Web. I will
point you to some of the better tutorials later on. I have taught workshops and a course on
searching the Web in the School of Computer Science at Carnegie Mellon. So I will list some
websites I’ve created to bring together helpful tools for you. You will see that our Dialog and
LexisNexis knowledge is a good foundation for retrieval with Web-based search tools.
Search engines are a big part of life here at Carnegie Mellon. The very first search engine,
Lycos, was created here by the scientist, Fuzzy Mauldin. Carnegie Mellon provided funding for
the project and lo and behold, the stock offerings from Lycos paid for a new building called
Newell-Simon, named in honor of two of the co-founders of artificial intelligence who taught
here: http://www.cmu.edu/tour/ (click on Walk Your Own Path and then click on bldg. #4.
FYI, I work in #7, Wean Hall.).
The most recent ventures that started here are Vivísimo and Yippy (formerly known as
Clusty). The co-founders knew the value of the skills of librarians. So they consulted with
many of us librarians at Carnegie Mellon numerous times while developing the product—now
only about 10 years old! Finally, because of this history and ongoing collaborative work at
Carnegie Mellon, Google opened up an office in Pittsburgh.
  Before proceeding, review the file: Week9.ppt and walk through the slides.
Search Tool Comments
One of the important distinctions most tutorials make is between “search engines” on the
one hand and classification schemes (Subject Categories and Directories) or specialized
collections (Virtual Libraries) on the other.
With search engines, you enter words (perhaps with logical operators) describing your
interest and hope for a match. Google, Bing and Excite are only a few examples. To do searches
in Alta Vista (like you have done on Dialog and LexisNexis), visit the Alta Vista home page
(http://www.altavista.com/), as well as the Advanced Search site
(http://www.altavista.com/help/search/help_adv). The latter will show you that Alta Vista
can handle Boolean and proximity operators, phrase searching, truncation, and its own varieties
of field searching. Many of the fields it handles are unique to the Web. For example, you can
restrict a search to terms that occur in the URL field. You can ask for all the pages that are
linked to a given page such as: http://www.cs.cmu.edu/~missy/. Google Advanced Search
provides a form to enable you to take advantage of various searching capabilities:
Week 9
INFO 522
2
http://www.google.com/advanced_search. Look over it closely for the many features that are
offered.
Virtual libraries are collections of annotated bibliographies on particular subjects. The
annotations generally contain clickable links to Web resources. The WWW Virtual Library is an
example of this: http://vlib.org/. Sometimes libraries may refer to these as resource pages.
With Subject Categories and Directories, you generally navigate through hierarchical term
displays, clicking on terms that catch your interest. Often you will move from general terms to
more specific terms as you go. Yahoo is an example of a directory system, a hierarchical subject
classification scheme. Yahoo provides a search engine, but its main claim to fame is this
directory scheme: http://dir.yahoo.com/. It functions much like the Dewey Decimal System
and is arguably no better (some would say it not as good). Web pages are assigned to subject
categories of the scheme by human beings, not software agents. But there are no logical
principles for creating or subdividing categories—they are simply slapped together to cover
whatever turns up on the Web.
To develop and maintain expertise in searching the Web, probably the best single site for
librarians to monitor is Search Engine Watch (http://searchenginewatch.com/). When you visit
the site, you will see that there are many search engines. Take my word for it that they are
NOT all equally good. They vary in coverage (the number of websites they give access to),
capabilities, and basic algorithms. They vary in their principles for ranking retrieved
documents by relevance. You can use Search Engine Watch to find comparative feature charts
and evaluations for the many competing engines on today’s market. A good principle to live by
in doing Web page retrieval is to use more than one search engine/directory. Most experts
tend to include Google and Yahoo.
What about the meta-search engines? They can search multiple search engines all at once—
sort of like a Dialog OneSearch. Is it a good idea to use them? Well, I tend to use them when I
want to confirm what, if anything, I may have missed in my search. In other words, I’m
checking for thoroughness in my search results. In such cases, the leading meta-search engine
is now Yippy, created by Vivísimo.
It is not necessarily a good thing to use meta-search engines that enter your search phrase
into many different engines for simultaneous retrievals. For one thing, inferior search engines
are often included. For another, search engines may differ in their requirements and so what is
a valid input for one may not be valid for another. For instance, you have probably learned to
put quotation marks around two or more words that you want a search engine to treat as a
phrase. But not all search engines will recognize that as a phrase search. Instead, they will
simply treat the words in the phrase as if they were ORed together, with highly unsatisfactory
results.
Week 9
INFO 522
3
Google Scholar
In 2004, an invention arrived that had a significant impact on the use of search engines AND
on libraries. It’s called Google Scholar. This tool enables you to search specifically for
scholarly literature. They state that the tool will find articles from a variety of academic
publishers, professional societies, preprint repositories and universities, as well as scholarly
articles available across the Web. Your search results include peer-reviewed papers, theses,
books, preprints, and technical reports from all broad areas of research.
Google Scholar analyzes and extracts citations and presents them as separate results, even if
the documents they refer to are not available online. This means your search results may
include citations of older works and seminal articles that appear only in books or other offline
publications.
Long before the announcement of Google Scholar, they had been working with various
publishers to deliver search results from certain databases when people search using Google.
Of course, the hitch (which they do NOT make self-evident) is that people can only see the fulltext of their search results IF their home institution (e.g., Drexel’s Hagerty Library) subscribes to
those particular services and/or databases. The three key examples of databases that have
made these arrangements with Google are: ACM Digital Library, IEEE Xplore, and OCLC
WorldCat.
My key concern is that students, especially undergraduates, have not learned in their
classes that this access does NOT come for free. Places like Drexel pay about $120,000 per
YEAR for access to IEEE Xplore—just one of numerous databases they subscribe to. So most
Drexel people do NOT realize that when they get Google results linking to hits from the above
three databases, they can view the articles because the Drexel libraries subscribe to those
services.
Jay Bhatt, the Engineering Librarian at Drexel, has been a critic of Google Scholar. He has
made postings to listservs where all of us librarians have discussed this tool. He pointed out his
concerns about engineering and science students relying too heavily on Google Scholar because
(I place my comments in blue):

It does not index e-books and handbooks such as those from EngNetBase and Knovel,
etc. (These are two leading engineering e-book services.) Also, Books24x7 and Safari are
important e-book services but not yet available in Google Scholar.

Google Scholar does not provide information on what is being covered, what journals
are indexed, what other databases are covered. So just relying on Google Scholar may
not be helpful. In other words, you are not necessarily retrieving the best possible hits
for your research. Worse yet, students may be misled to think that if they found little,
they’ve exhausted all of their possibilities.
Let me explain this a different way. If a librarian was helping a student find journal
articles on a computer engineering topic, they would want to ensure that the student
searched INSPEC (some of the full-text articles found in INSPEC are available in IEEE
Xplore), Compendex, and Web of Science. We could throw in ACM Digital Library for
Week 9
INFO 522
4
good measure. If a student did not search ALL of these, then they would be potentially
missing out on important citations that could prove crucial to their research.
So I encourage you take a look at Google Scholar (http://scholar.google.com/) and use it to
find a FEW of the citations you will list in your final project. Note that I say a FEW because in
the assignment I also caution you to use Web resources with care. There are many excellent
items on the Web. But there is also a lot of junk. Look for a respectable level of scholarship if
you decide to include a few Web-based materials and apply the principles for evaluating
websites that are contained in one of the readings below.
In our readings, I will offer a couple of articles for you to learn more about this tool. Then
you can decide for yourself if you think it’s worthwhile or a lot of hype or a combination of
both.
WolframAlpha
In 2009, a new model of searching was launched called WolframAlpha. If you’re unfamiliar
with the name, Stephen Wolfram is the founder of Wolfram Research. He is a British physicist
and mathematician who was once attached to the famous Institute for Advanced Study in
Princeton, NJ. That’s the same institute where Einstein once researched. In 1988 he launched a
computer algebra system called Mathematica that has significantly impacted the worlds of math
and computer science.
In 2002, Wolfram published a rather contentious book, A New Kind of Science (NKS), which
presents an empirical study of very simple computational systems. He argues that these types
of systems, rather than traditional mathematics, are needed to model and understand
complexity in nature. His conclusion is that the universe is digital in its nature, and runs on
fundamental laws which can be described as simple programs: cellular automata. He predicts a
realization of this within the scientific communities will have a major and revolutionary
influence on physics, chemistry and biology and the majority of the scientific areas in general,
which is the reason for the book’s title.
Wolfram|Alpha is being touted as a computational knowledge engine with a new approach
to knowledge extraction. The engine is based on natural language processing, a large library of
algorithms and an NKS approach to answering questions. The new engine differs from
traditional search engines because it does not simply return a list of results based on a query—
instead it computes an answer. Some industry observers, such as Nova Spivack, claims that “it
could be as important as Google.”
Readings
REQUIRED

Hagerty Library, Drexel University. (2004). Evaluating information on the Web.
http://www.library.drexel.edu/documents/tutorials/webeval/intro.html
Week 9
INFO 522

Meola, M. (2004). Chucking the checklist: A contextual approach to teaching
undergraduates web-site evaluation. Portal: Libraries and the Academy, 4(3), 331-344.
(Available through Summon at Hagerty Library.)

Tenopir, C. (2005). Google in the academic library. Library Journal.
http://www.libraryjournal.com/article/CA498868.html (very slow to open up)

Kolowich, S. (2010). Searching for better research habits. Inside Higher Ed.
http://www.insidehighered.com/news/2010/09/29/search
5
This article touches on the challenges faced by librarians today, especially academic
librarians. After reading the article, make a point of looking at the comments at the end
as well.

Price, G. (2009). Great day for power searchers: Google adds new search options.
ResourceShelf. http://www.resourceshelf.com/2009/10/01/google-search-adds-newfeatures/

University at Albany, University Libraries. (2010). The Deep Web.
http://www.internettutorials.net/deepweb.asp

UC Berkeley. (2010). The Invisible Web: What is it, why it exists, how to find it, and its
inherent ambiguity.
http://www.lib.berkeley.edu/TeachingLib/Guides/Internet/InvisibleWeb.html

Sutter, J.D. (2009). New search engines aspire to supplement Google. CNN.com.
http://www.cnn.com/2009/TECH/05/12/future.search.engine/index.html?iref=news
search
Recommended

Price, G. (2003). What Google teaches us that has nothing to do with searching.
Searcher, 11(10). http://www.infotoday.com/searcher/nov03/price.shtml

Abilock, D. (2010). Choose the best search engine for your information need.
http://www.noodletools.com/debbie/literacies/information/5locate/adviceengine.ht
ml

University at Albany, University Libraries. (2010). How to choose a search tool.
http://www.internettutorials.net/choose.asp
Assignment and Discussion Board

For this week, I am placing practice exercises AND a new assignment in the assignment
folder. Only turn in the assignment, NOT the practice exercises.

Submit your assignment by Midnight, Sunday, March 6th (Eastern time).
Week 9

INFO 522
6
You are required to make some kind of a discussion board posting for the week. Please
go to the Discussion Boards to respond to the topic for this week. Complete your
discussions by Midnight, Sunday, March 6th (Eastern time)
 Submit questions or comments about the challenges and issues you face as you try to
search using various Web tools.
Download