Montana State University
on the net
Tracking Title Search Capabilities
ONLINE, May 2001
|Web page title search capabilities have expanded somewhat, although to nowhere near the levels available in our more sophisticated bibliographic systems.|
Online library catalogs and periodical indexes now have a wide variety of title search capabilities. Titles can be browsed in an alphabetical list. Keyword searches for single words within a title are available along with exact title and title phrase searches. And bibliographic databases can even have multiple title fields for main title, subtitle, series title, alternate title, uniform title, and more.
On Web pages, there is only one title, designated by the HTML title element. Title searching on the Web goes back to the earliest days of the Web search engines. As the early HTML documents began using the title element, it became an important element for early search engine ranking algorithms. AltaVista even made it searchable. Since that time, Web page title search capabilities have expanded somewhat, although to nowhere near the levels available in our more sophisticated bibliographic systems.
|An ideal system would recognize when a title phrase match works better than a title word match.|
furry snake lunar
snake amazes tourists
furry snake amazes lunar tourists
furry snake lunar
snake amazes tourists
In the first example, a title phrase search where the record that is an exact match for those words in exactly that order would be the right match. For the second, any title that matches the phrase at the beginning of the title might be the correct match. For the last two, a title word search works best, where the hits all contain those words in the title somewhere, but not necessarily in that order.
An ideal system would recognize when a title phrase match works better than a title word match. However, ideal systems do not yet exist. For those search systems that provide both, the searcher needs to choose which to use. One-word titles are problematic, since a title word search for Science finds every title containing "science." Library systems may suggest using a title phrase search for single word titles. While that works, it is counterintuitive to search a single word as a "title phrase."
Even so, for Web pages that do use titles, and use them appropriately, the title is often a good summary of what the page is about. With the lack of any authoritative and consistently applied subject indexing of Web pages, the titles provide one of the best subject access points for searchers. A page entitled "Market Share in the Light Bulb Industry" is likely to have information on market share in that industry whereas a page that only contains those words in the text may simply be a collection of jokes. Title searching can help get better retrieval for searches for organizations, biographies, lesson plans, FAQ pages, and other subject-oriented needs.
|While the standards community found agreement in the CCL and a few online vendors adopted (or adapted) it, there is still a wide range of ways to search titles.|
Not so long ago, the Common Command Language (CCL), now an official standard, was the rallying cry of the information industry user. Why not have Dialog, SilverPlatter, OCLC, library catalog vendors, and all the others use a common syntax for special searches such as the title search? Now in the bibliographic realm, two main kinds of title searches were considered. Title keyword searches looked for the search terms appearing anywhere in the title. A title phrase search would look for an exact match of the whole title.
While the standards community found agreement in the CCL and a few online vendors adopted (or adapted) it, there is still a wide range of ways to search titles. We can use /ti on Classic Dialog and STN. SilverPlatter likes the in ti while EBSCO Host and InfoTrac use just ti followed by the search terms. While ProQuest offers TI(), FirstSearch uses ti: for title keyword searching and ti= for a title phrase search. The one common element is the "ti" abbreviation for title, which all of the Web search engines have avoided in favor of the full word "title."
For the advanced searcher, AltaVista's original approach of field name, colon, query term is available from several search engines. On AltaVista, a search for "market" in the title can be entered as
title: market. Note the use of "title" rather than "ti."
Multiple word title field searches are where the bibliographic database separation of title keyword from title phrase searches meets the Web. Which search engines can handle title word searching and which do title phrase searching? Can a space be used after the colon?
First of all, which even support title searching? AltaVista, AlltheWeb, Northern Light, Lycos, some Inktomi partners, and Google all have some capability for title searching. All but a few of the Inktomi partners have title searching available via a form with a drop-down list where a title search is one option. In addition, all but Google support the
title: command line syntax. Google chose a different path.
Using that form option, Google does not give the searcher a chance to combine a title search with terms that would be elsewhere on the page. The command line version of the scripted ability on the Advanced Search page is the
allintitle: field search, which requires all the terms to be in the title but in no particular order. However,
allintitle: cannot be used in combination with other search terms.
The other title search option on Google is not available on the Advanced Search page but only on their regular search. This is the
intitle: field search that can be used in combination with other terms that need not be in the title. However,
intitle: can be used for only one word at a time.
So Google has added some title searching, using two different commands, but they each have their limitations. And both function a bit differ- ent from the more usual
title: command that other search engines use.
intitle:only handles one word, what happens for multiple words entered after the colon on the other search engines? A search for
title:green tomato pie
will usually search only for "green" in the title. AlltheWeb, Lycos, the Inktomi partners, and Google's intitle: will process that search as
(title:green) AND tomato AND pie
while AltaVista processes
(title:green) OR tomato OR pie
even though AltaVista's relevance ranking brings records whose titles contain all three terms to the top. Only Northern Light searches for records where all three terms appear in the title.
All the search engines discussed here support phrase searching using double quotes. If a searcher wants to try the above search as an exact phrase search, the typical syntax is
title:"green tomato pie"
AltaVista, AlltheWeb, Lycos, Northern Light, and Google (using
allintitle:) support that syntax and retrieve only records that contain that exact phrase within their title.
Unfortunately, most of the Inktomi partners (such as HotBot, MSN Search, and iWon) choke on that syntax and end up processing it as if "title" were a search term:
title AND "green tomato pie"
HotBot could do a phrase search in the title if, instead of the title: syntax, the phrase is entered in double quotes in the search box and the drop-down "Look for" menu choice is changed to "the page title." However, using that same strategy on MSN Advanced Search fails to work properly. So check carefully when trying to search a title phrase with an Inktomi partner.
title:syntax, the search query should come right after the colon with no space. For example, a search for the word "green" in the title would be
If there is a space after the colon, as in title: green, then AlltheWeb, Lycos, Google (using
intitle:), and the Inktomi partners will not process it properly. Instead, they will consider "title" to be a search term. AltaVista, Northern Light, and Google's
allintitle: work just as well with the space after the colon. Strangely enough, AlltheWeb and Lycos can handle a space after the colon if a phrase is searched. But since they will all work without the space, just leave the space out.
In general, using the form or drop-down menu options can be effective ways to search titles, especially for simple title searches. However, combining a title search with additional search terms can only be done via the forms at AlltheWeb, Lycos, Northern Light, and HotBot.
While title searches can be very helpful for the advanced searcher, the current implementation at each search engine is rather confusing. Learn the details of one or two, but be wary as their capabilities may change at any time. For a quick overview, check my title search page (http://www.searchengineshowdown.com/features/title), which I will try to keep up-to-date as the title search features change. And then we can all try to figure out how to teach one syntax for use in our bibliographic databases and another for use on the Web.
Greg R. Notess (firstname.lastname@example.org; http://www.notess.com/) is a Reference Librarian at Montana State University.
Comments? Email letters to the Editor at email@example.com.
Copyright © 2001, Information Today, Inc. All rights reserved.