Instructions for CD Search Engine

The search performs what is called an "AND" search, meaning it searches for each file which contains every word listed. For example, if a file contains three of four words queried, it will not appear. Type words in any order.

The search will match words as substrings. In other words, if you type "viol" it will also match files which contain the word "viols" without necessarily the word "viol" itself. The same holds for other, perhaps less obvious, combinations.

Along with all of the CD files, this search will look in the related discography files in either the composers or performers section of the FAQ. For non-discographic information, aside from the notes of historical explanation in the CD files themselves, it is suggested that you try either a topic from the FAQ front page or the link search to find linked resources.

The search engine has now been updated to search only actual CD files (or related, recently-available media like cassettes) by default. This is due in part to the increased volume of LP or even older media which are now on file. To search all of these media, some dating to the early part of this century, check the "include other media" box and the search will be returned in two parts. LPs or other media which have been reissued on CD should appear in the CD section of the search results, and there is no need to check the "include other media" box to find them.

Results are returned in order of precedence, based on a simple "scoring" algorithm. In this way, hopefully the most desirable files will appear near the beginning of the list of results. This is a new feature of the search and will hopefully prove worthwhile.

The search is not case sensitive, so that either upper or lower case letters are sufficient. Accented & special characters will hopefully work properly, so that either "pérotin" or "perotin" will match the same files, although there may be some problems there with differing browser character sets. If there is a problem with an accented character, it is advised to use the unaccented version, as either should match.

Words which appear too often in the CD files will be ignored by the search engine, as will all words shorter than three characters. If your query contains only words of this type, an error message will be returned. If not, such words will be ignored silently.

Further refinements of the searching process can be obtained by using the "refined" search, featuring regular expressions and requiring a more advanced knowledge to use effectively. This is a powerful tool, but unnecessary for the vast majority of queries, although it does provide an easy way to do exact searches of words in the given order.

Please do not abuse this search engine. I have to pay for it!

For the software-minded, this is a hashed, caching search engine written by Todd McComb in 1998. Its internal databases are updated incrementally, once per day.


To CD index page

To Early Music FAQ

Todd M. McComb