On the Net, The Death of Open Text and the Birth of Livelink Pinstripe

Greg R. Notess
Reference Librarian
Montana State University

ON THE NET

The Death of Open Text and the Birth of Livelink Pinstripe

DATABASE, August 1998
Copyright © Online Inc.

In terms of a database, the primary offering from Livelink Pinstripe is its database of Web sites.

Where do Web search engines go to die? An engine's database of Web sites may be valuable at first, but due to the dynamic nature of the Web, it requires constant updating. More importantly, on the Internet, the market share of the search engine and its user base are not to be lightly squandered. Some of the early Web search engines have certainly fallen behind their newer competitors. The World Wide Web Worm (WWWW), an early and popular search engine that still attracts searchers, has now metamorphosed into GoTo.

Open Text Index was another early search engine. As the first Web search engine to offer field searching, Open Text led the way to more advanced search features. Unfortunately, its database was relatively small. Other competitors, such as Lycos, Excite, and Infoseek, were busy adding subject channels, news, and supplementary databases. The Open Text team took a different approach.

As a general Web search engine, the Open Text Index is dead. It is no longer available at Open Text's page, and the URL now points to a new service. This is a new Open Text initiative: Livelink Pinstripe. Rather than trying to build a bigger database of Web sites or compete directly with the others, Open Text chose a subject-specific route. Livelink Pinstripe, designed specifically for business users, is a database of business information and sites available on the Web.

Livelink Intranet is the name of Open Text's knowledge management application for intranets. While Livelink Pinstripe carries the Livelink name, it is a separate product that aims to show the capabilities of Livelink Intranet as well as providing a useful database for Livelink Intranet customers. Fortunately, it is also free for the entire Web community.

WHAT'S IN LIVELINK PINSTRIPE

... the Pinstripe database cannot be considered a comprehensive database of business Web sites, although it can bring up a higher proporation of relevant hits on some searches.

The overall structure of Livelink Pinstripe includes seven main categories outlined on the graphics bar near the top of the page. The Search section is the default starting point. The other sections are Finance, News, Employment, Companies/People, Travel, and Discussion. Each of these sections, and their subsidiary pages, have the standard banner ads, but they are no more intrusive than most other search engines' banner ads. In these early stages, many are simply ads for Livelink Intranet.

The main Pinstripe database is available under the default Search section. The other six areas consist primarily of links to other companies' databases and services. In terms of a database, the primary offering from Livelink Pinstripe is its database of Web sites. It is similar to, but certainly smaller than, the other general Web search engines' databases. Open Text is not accepting submissions, at least at this time. Instead, the documentation states that Open Text is "being very careful to build this database gradually, ensuring that it stays focused on business." It uses lists of URLs from "reliable business sources," which include business magazines. However, it also mentions that it includes the Web sites of Livelink Intranet customers. Since it is still actively being built, and the documentation does not go into much depth on the issue, the scope of its content must be determined by what can be found from searching or browsing the database.

Pinstripe appears to consist mostly of commercial sites, those with a .com ending. There are remarkably few government sites currently available in the database. A Power Search on gov in the URL only brings up about 500 hits, most from www.eletrobras.gov.br (the Brazilian electric company about to be privatized) and other non-U.S. government Web sites. A few nasa.gov sites show up, but none from primary government Web sites with business information such as sec.gov, census.gov, doc.gov, or bls.gov. There are more .org and .edu sites, but it is still a very limited number of pages from non-.com domains.

Moreover, the database certainly does not contain all the .com addresses. It is only a limited subset of commercial sites, selected from their "reliable business sources." This excludes many Web pages that can clog up a search of one of the large, general Web search engines: adult sites, personal pages, and fan pages. It succeeds well in eliminating some of the junk, but it can also be difficult to find pages in Pinstripe mentioning smaller companies or very specific business topics. The result of this approach is that the Pinstripe database cannot be considered a comprehensive database of business Web sites, although it can bring up a higher proportion of relevant hits on some searches.

One advantage to having a smaller database is that it can be refreshed more frequently. According to the documentation, the entire Pinstripe database is refreshed every two weeks. If they can deliver on the frequency, there should be very few dead-end links in their database.

SEARCH OPTIONS

The main search section of Pinstripe presents three different search options: Slice Search, Quick Search, and Power Search. The default among these three is the Slice Search. This option enables users to just search "slices" of the database, predefined industry groupings arranged in hierarchies. There are more than 150 slices, although only the top twenty or so hierarchical categories show up on the initial screen. A search can be limited to one of those broader industries, or by clicking on one of them, the subsidiary slices can be used to limit the search. Unfortunately, multiple slices cannot be chosen.

These slices are determined during the building of the index. Individual pages are placed into one of these specific slices based on a determination made by the indexing agent. Thus, this is automated indexing similar in concept to what Northern Light does with its index, although on a much smaller scale. While using these slices as limits can reduce the number of search results, do not rely on them to be comprehensive or completely accurate. Looking under the slice of Mining, five hits came up for gold mining, but when searching all slices, 42 records were found, many of which were relevant.

If no slice is chosen, the default will cover all sections. The same number of hits will result from a single term searched in a Slice Search with the default of all as in a Quick Search or a Power Search. The Quick Search option simply removes the display of slices. It also links to a few other searchable databases from Pinstripe Partners. The Power Search provides more search features and looks just like the old Open Text Index's Power Search.

In the Power Search, multiple words can be combined with AND, OR, BUT NOT, FOLLOWED BY, or NEAR. FOLLOWED BY requires that the second search term follows within 80 characters of the first, while NEAR means that the two search terms occur anywhere within 80 characters of each other.

The terms can be searched anywhere on a page or in any of the following fields: title, first heading, summary, and URL. The summary is a combination of the title of the page, its first heading, and some "important" text extracted from the page. The title field is the HTML <title> element. The first heading searches any text surrounded by the HTML <H1> tags. The URL field includes any portion of a page's URL separated by punctuation marks.

Multiple terms entered in any of these three search options are searched as phrases. No quotation marks are needed, although it does not cause a problem to add them. In all three options, Pinstripe presents the results ten at a time. Each hit has a title, file size, URL, and a summary of the Web page. No sorting of the results is available. However, at the top of each page is an opportunity to further refine the search by adding new search terms with any of the operators and/or in any of the available fields.

Finding search help and documentation is a bit difficult in the current incarnation of Pinstripe. On the top graphic bar, there is a drop down menu that reads Go To ... by default. Click on the menu to see a link to the Livelink Pinstripe Help page.

FINANCE, NEWS, AND EMPLOYMENT

And what about the other sections of Livelink Pinstripe? The second level of the top graphic bar gives tabs for easy movement to other sections. The first three after Search are Finance, News, and Employment. Other than the search section, all of these other areas are moving beyond the Pinstripe database to point to other selected Web sites and direct search connections to other companies' databases. The Finance section brings up a search form for Quote.com stock quotes along with selected links to other stock resources, such as Dow Jones Online, PCQuote, the Wall Street Journal, and Motley Fool. For more finance information, the area has tabs for going to new pages for Portfolio Tracking, Mutual Funds, and Company Research. The Company Research tab provides a quick search from Zacks Investment Research, portfolio tracking has a PCQuote search form, and the mutual funds tab offers searches from both Forbes Mutual Fund Lookup and Globefund Fund Selector.

The news section uses the Open Text/Livelink Pinstripe search engine to search several business periodical Web sites including Fortune, Forbes, Financial Times, San Jose Mercury News, Investor's Business Daily, Business Week, Bloomberg, USA Today, Worth, and Toronto Star. For all of these sources, the search only points to freely available pages. The news section also has another tab for Forbes Daily, which provides several headlines from the Forbes Daily Web site along with links to other sections of the Forbes site and to the stories themselves.

For the Employment portion of Pinstripe, the default link is a Net- Temps U.S. Job Search form. In addition to this National Job Search, other tabs lead to Jobs by City, Submit Resume, and Locate Staffing Firms.

COMPANIES AND PEOPLE

... use Companies Online with caution, as it appears to be a rather roughly amalgameted collection of data.

In the Companies/People section of Pinstripe, the default page brings up a searchable form labeled Companies Online. It provides Web site URL, email address, company contact information, and a brief description of the company and its products. However, this is WhoWhere's Companies Online, a directory of companies with an online presence, rather than the well-known CompaniesOnline site (http://www.companiesonline.com), reviewed in DATABASE, December 1997, pp. 30-32. Much of the information appears to come from the Whois database of Internet domain name owners. Some data appears to be generated by matching a yellow page database to the Whois database.

While it claims over 500,000 entries, compared to the 100,000 in Companies Online, the records are a bit messy. For example, a local Kinkos appears in the database with a local address and phone number, but the URL given is the one for the entire chain. Another very small local company had contact information but the URL for Netscape. Many publicly traded companies will also link to stock quotes and EDGAR filings, but use Companies Online with caution, as it appears to be a rather roughly amalgamated collection of data.

The other tabs on the Companies/ People section are for Email Addresses, Phone & Address, Yellow Pages, and Toll Free Numbers. All of these are searchable connections to various parts of WhoWhere's databases. On each page, there is also a short list of other directory databases on the Web.

TRAVEL AND DISCUSSION

The Travel section begins with a Travel Web Airline Search. Other tabs include Cars, Hotels, Maps, Weather, and a Currency Converter. Each section includes selected links pointing to the main Web sites for rental car companies, hotels, and other relevant links. The discussion tab presents a DejaNews search form for finding Usenet news articles and a Columnists page. The latter links to selected business columnists including Steve Harmon from the Internet Stock Report, Tom Petruno of the Los Angeles Times, and Andrew Serwer of Fortune.

While all of these extra sections offer access to useful Web resources, there are plenty of other sites that provide similar pointers, search forms, or at least links to comparable resources. What sets Livelink Pinstripe apart is its searchable database of Web sites that tries to limit itself to sites of interest to the business community. It could use more reliable response time and a larger database, especially in its coverage of government and other noncommercial Web sites with substantial information of interest to the business community. While the database currently seems too limited for many business uses, try it out for your own business searches. As Open Text continues to build the database, Livelink Pinstripe may well find its niche in the business community. At the least, it should prove an incentive for Livelink Intranet customers.

The Open Text Index had some unique features. And while it is now dead, Livelink Pinstripe has risen from its ashes. By taking on the Power Search features of the Open Text Index and targeting a specific user market, it may well be setting a new model for the development of Web databases.

Communications to the author should be addressed to Greg R. Notess, Montana State University Libraries, Bozeman, MT 59717-0332; 406/994-6563; greg@notess.com ; http://www.notess.com.