The learning
Once it is understood that search engines rank documents based on a specific understanding of the way the Internet functions, it then follows that in order to insure that new document types and technologies are able to be read and that the algorithm be changed as new understandings of the functionality of the Internet are uncovered a search engine must have the ability to “learn”.
Aside from a search engine needing the ability to properly spider documents stored in newer technologies, search engines must also have the ability to detect and accurately penalize spam and as well as accurately rank websites based on new understandings of the way documents are organized and links arranged. Examples of areas where search engines must learn in an ongoing basis include but are most certainly not limited to:
Understanding the relevancy of the content between sites where a link is found
Attaining the ability to view the content on documents contained within new technologies such as database types, Flash, etc.
Understanding the various methods used to hide text, links, etc. in order to penalize sites engaging in these tactics
Learning from current results and any shortcoming in them, what tweaks to current algorithms or what additional considerations must be taken into account to improve the relevancy of the results in the future.
The learning of a search engine generally comes from the uber-geeks hired by and the users of the search engines. Once a factor is taken into account and programmed into the algorithm it them moves into the “knowing” category until the next round of updates.
|