|
|
The need for "Research Quality" search engines
Opportunities for improvement -- Understanding of search
Filtering and accuracy
Reflect on this perspective: The fastest way to find something is to determine first where it is not. That may seem a small point, but it makes all the difference in the design of a search engine. The best search engines act in part as filters; they quickly disqualify every record where one or more desired terms is missing. This filtering process is essential to accuracy. With well structured indexes the process can be carried out at blistering speed.
The pity is that this perspective dates back over 20 years. Filtering technology was built into at least one search engine that arrived in the marketplace back in 1985. Why haven't the Internet engines picked up on it?
Efficiency
Each false positive among the list of hits is evidence that a search engine is less accurate and less efficient than it could be. If you could "lift the hood", so to speak, on such a search engine you would find glaring inefficiencies in its technique and (probably) in the structure of the indexes that it is using. When a major search engine offers 8 million hits with false positives near the top of the list among the (supposedly) most relevant hits, that engine is a prime candidate for technical improvement. Yes, it will produce fewer hits. That's okay. A few good hits always trump millions of bad ones. And more good news: If they filter first, the search engines will actually run faster.
Garbage elimination
Search is about garbage elimination, not garbage presentation.
Previous | Research Quality | Next
|
|||||
| words close together.com | The "Research Quality" Search Engine by Marpex, Inc. | ||||