By Jennifer Mills

Through the magic of the World Wide Web, a vast electronic ocean of software development information is available at the mere touch of a computer keyboard. This spring, BASIS has cast a few more megabytes of important information into the Internet ocean with the premier of the BASIS Online Technical Support Knowledge Base at www.basis.com/support/kb/. Now developers can access the same extensive library of product information and troubleshooting tips used by BASIS' own technical support analysts any time of the day.

What Is An Online Knowledge Base?

An online knowledge base is a library of short technical articles, typically accessible through a search engine front end. Unlike a white paper or general article that provides an overview of a technology or a product, the information in an online knowledge base is organized to provide users with detailed answers to very specific technical questions. For example, a developer that wants to find out more about the overall workings of the BASIS ODBC Driver™ would find a white paper most helpful. However, when this developer needs to know precisely how to solve an ERROR 61 with the ODBC Driver, he or she would want to search an online knowledge base.

The BASIS Online Knowledge Base is comprised of over 250 articles put together by BASIS' own technical support analysts. The articles cover every BASIS product, many common BBx and operating system error messages, and address the most frequently asked questions from customers. In order to keep the BASIS Knowledge Base targeted and easier to use, it does not include white papers or product demonstrations, but does incorporate relevant BASIS Advantage articles.

Simple Searching

BASIS' Online Knowledge Base uses a search engine by Excite as its front end. Like its Internet cousin, this knowledge base version of the Excite engine accepts a query from the user, searches the online library of articles, and returns an index of relevant articles.

Because the Excite engine is a concept-based searching engine, simple queries will pull up an index of up to fifty articles that include the words requested in the search, as well as articles that cover similar concepts. This feature helps ensure that users will get a wide variety of resources to examine with each individual search. For example, if you request articles on SQL, the engine may also pull up relevant articles on an ODBC Driver.

Advanced Searching

For general questions, a simple query can provide you with a wide variety of information to examine. However, when you need to answer a very specific question, a simple search may provide you with too much information. In these circumstances, a more focused or advanced search of the knowledge base may be needed.

One simple way to narrow down a query is to use more words. If a user wanted some specific information on how the BASIS ODBC driver works with Windows NT instead of just entering ODBC Driver or Windows NT, the user could input ODBC or Windows NT and fewer, more relevant articles may appear in the resulting index.

Another more sophisticated method of searching uses Boolean logic operators to tell the search engine what words must or must not be included in articles it returns. There are eight different words and symbols that can be used to focus a search. Remember that when doing a Boolean search, the Boolean operators must be completely capitalized but the keywords themselves are not case sensitive. Also note that when using the Boolean symbols +, -, or *, no spaces should be placed between the keywords and the symbols.

Boolean Operator Table
AND or +	Putting a + or an AND in front of a word ensures that the search engine returns only articles including that word. For example, the query data + server or data AND server will return articles that include both these words.
NOT or -	Putting a - or a NOT in front of a word ensure that none of the articles returned contains that word. For example, the query ODBC-16 bit or ODBC NOT 16 bit or ODBC AND NOT 16-bit will prevent the search engine from bringing up any articles relating to 16-bit ODBC drivers. Be aware that in the process of excluding the reference of 16-bit in any of the articles mentioned, you may also have eliminated any articles that contain both 32-bit and 16-bit references.
OR	Putting an OR between two words ensures that the articles pulled up by the search engine include at least one of the two words. For example, to find documents that have either the words ODBC or SQL you would enter ODBC OR SQL.
()	Parentheses can be used to group large sections of a Boolean search together. For example, to find articles that have the word SQL, and either the word PRO/5 or BASIS ODBC Driver, you could enter SQL AND (PRO/5 OR ODBC).
*	The *, known as a wildcard symbol, will return variations of a word when only the first few letters are given. For example, a search for config* will return articles with config, configure, configuration, config.bbx, configuring, etc.

Grouping By Confidence Or Subject

The Excite search engine offers a subject/confidence grouping option that lets users decide how articles are indexed and returned to them after a successful search. When grouped by subject, articles returned from a search will be indexed according to how many times a search word appears in a particular article. When grouped by confidence, the search engine will return an article index organized by a "confidence factor" percentage. This confidence factor reflects how well an article addresses the subject of the original query. Here at BASIS, we have found that grouping by confidence consistently results in more effective searches.

Technical Support Still A Phone Call Away

Technical support analysts are continually adding new articles to this online library in order to keep the knowledge base current and provide the best information to programmers. However, even the best knowledge base cannot replace a live technical support analyst. That is why BASIS will continue to support developers directly via the phone at 505.345.5021, through faxes to 505.344.9057, and by email at support@basis.com. BASIS' Technical Support analysts also encourage developers to report any inaccuracies in the Knowledge Base directly to them. Customers experiencing any problems with the BASIS Online Knowledge Base can contact BASIS'online services coordinator for help at webmaster@basis.com.

Note: the next article includes a step-by-step example showing how to use the Knowledge Base search engine to find information on a particular topic of interest.