WebCompass shows the way to Internet resources
Ever been trolling the Web in search of some obscure piece of information only to be frustrated by inconsistent search results delivered by services such as Yahoo, Lycos and Infoseek?
Quarterdeck Corp. has an answer. Its WebCompass utility lets you conduct searches using multiple engines simultaneously and delivers all the results in one fell swoop. Properly configured, we found this utility to be a super time-saver.
WebCompass runs as a local search application that can be called from any Web browser. It queries a customizable list of Internet search engines, which it calls resources, and returns all their results to a single screen. Users can then save the results of a query and have WebCompass index and update the results automatically by periodically searching again without user intervention.
This is useful in garnering the latest information on a particular topic – for example, if you want to keep a current list of all the newest sites that offer RealAudio feeds.
The first step toward finding your way with WebCompass is installing its components. The installation routine installs Quarterdeck WebServer and, optionally, Quarterdeck WinSock and Quarterdeck Mosaic as well as WebCompass. Via its Common Gateway Interface (CGI), Quarterdeck WebServer processes all queries and passes them off to the various search engines.
While you can use any browser with WebCompass, only Quarterdeck Mosaic will start automatically when you click on WebCompass. To work with others, you must manually start the WebServer and your browser, then point the browser to a WebCompass home page, which is a document on your local machine. We successfully configured WebCompass to work with Netscape Communications Corp.’s Navigator and Navigator Gold 2.0 beta.
We also customized the product to add additional search engines. The product comes with the search engines listed in the table above, but we added such useful engines as Infoseek and Alta Vista.
After you’ve added a search engine, you can test it to make sure you’ve configured it correctly before you add it to the complete list. This process simply performs a regular search of just the new engine from within WebCompass.
Launching the probe
Finally, we were ready to search. We specified a single-word topic in a single field, though multiple-word topics work equally well, and clicked on the Search button. In less time than we’ve waited for a single-engine search, WebCompass returned results from all the search engines on our list. Its speed was truly impressive.
Some of the search engines timed-out before they were able to return any results. You can set parameters within the product that control the amount of time the agent tries to query documents or resources.
Once we retrieved our documents, we could look them over using normal hypertext pointers. With most browsers, that’s where the functionality ends. But with WebCompass, you can mark specific documents and save them in what Quarterdeck calls a topic (see screen shot, page 44). Topics are special collections based on a keyword. They are automatically indexed when created.
The indexing process follows the links found by the search engines and retrieves the associated documents. It categorizes retrieved documents when the query is ambiguous and removes documents that don’t seem to fit in with what you wanted.
For example, we queried for “Pomeranian” and got back not only pages about our favorite toy breed of dog, but also geopolitical data about Pomerania in Europe. The indexing process then threw out the documents we would have thrown out manually, but it did so while we were working on other tasks.
An indexed topic can be activated, after which WebCompass’ agent will periodically (by default daily) query the Internet for new references and check links to documents within topics. This helps keep you up-to-date on your areas of interest without forcing you to take the time to search.
Once you’ve created a topic, you can search it again using another keyword to narrow your results.
For all of its utility, the product has a few drawbacks. Accuracy is the most critical.
When we told WebCompass to find “Gabrielle Reece,” it reported no hits on Lycos or OpenText. When we went in manually, we got more Reece’s pieces than we could digest – nine, in fact, with a relevancy rating of greater than 0.5 on Lycos and two with ratings of greater than 100 on OpenText.
On the other hand, there’s a lot of duplication among the various search engines, so what one fails to find is probably present in the results from another.
Another significant drawback is that, except for searching, the product is slow, even on our 100-MHz Pentium with 16M bytes of memory. Moving to an existing topic stored on the local hard drive takes as long as fetching a page from the ‘Net.
This appears to be because of overhead imposed by the CGI of the local Web server, where operations are interpreted before being acted upon. However, Quarterdeck has no viable alternative yet if it wants the product to work with any Web browser.
Another gripe is that the user interface requires too many clicks to move around. For example, after viewing all the documents retrieved in a search, you must click on each to add it to a topic or you can click a single box to get all documents from a resource. But there’s no box to click to get all documents from all resources.
After the selection, you must return to the top of the search page to create a topic; a second button at the bottom of the window would be a handy addition.
We’d also like a way of viewing the pages in a single list rather than having them sorted by the service from which they came.
In addition, we could find no way to simply delete a document from a topic once it was created.
Another issue is WebCompass’ documentation, which, at 24 pages total, is too skimpy. The product is simple enough to use, but we would have preferred to see more details about how WebCompass interacts with the browser and Web server it invokes.
Nevertheless, we expect WebCompass to become an essential business tool. The bottom line is that it saves time by taking on tedious tasks, freeing up users for more productive pursuits.
