Techniques for Accessing Data
Given our flat-file structure, Cat's Pajamas is about as open as a system can get. For industry-standard access using SQL an ODBC interface is easily crafted – our ODBC data dictionary facilitates the process.
Cat’s Pajamas pioneered this search method in a commercial system and we’ve steadily expanded usage. Cat's Pajamas became accidentally trendy—these search techniques are quite similar to those used in browser search engines. The power of this kind of search is best illustrated by the following example of a question from marketing or sales staff.
Can I see all orders in Oregon (state OR) for books on Coastal Fauna (subject CF) by park gift shops (customer type code PG) resulting from our summer sale (source code SU)?
This gets complicated because the search argument involves properties of the sale, properties of the customer, and properties of the book. For this example let’s assume there are 2 million line items in the file and that 25 records match the search argument. In a Cat’s system running on a modest Pentium computer, the 25 items will be on the screen in less than 5 seconds.
This search technique involves building a very large matrix that associates all combinations of data. It requires use of exact pointers to data locations, which is a determinant of our database foundation in fixed-length flat-ASCII files (see database notes below). This search technique is a particular strength of the DB/C programming language.
What system feature is used more often than any other in daily operations? The answer is searching. -Search techniques are used more frequently than any other function. The flexibility of search arguments and speedy response with matching data is a huge determinant of productivity. Cat's Pajamas is very strong in this area – we’ve seen jaws drop as untrained users find customers or books or sales information in huge databases within seconds. Efficient search capabilities yield bottom-line results. An example is a test conducted by a publisher to compare Cat's Pajamas to system “X.” After training, two operators entered the same orders in Cat’s Pajamas and in system “X.” System “X’ required three time longer.
Search for a customer in Cat's Pajamas using fragments of the name. For example, search
on “oop” and “light” to find “Hoople Enlightenment Center.” The speed is
impressive. If only five customers out of 200,000 have matching information, 1-2
seconds is all it takes to get them on screen. Address and city can be
incorporated in this search if desired. The beauty of this search is that the
more arguments you add, the quicker the results. The flexibility of this search
is essential for prevention of duplicate accounts.
See these other system design topics: