The special character tokens: ö, Ö, å, Å, ä and Ä are substituted with the corresponding character ÖÅÄ before they are added to the database.
Other characters like -,!:() are removed from the word before they are added.
For example, the word "Mac-lab!" causes the single words "MAC" and "LAB" to be inserted separatly.
Averge search takes about 0.3 sek, exclusive transfer time.
The Perl regular expression version is written i Perl.
> and < is replaced by > and <
Characters like åäö are substituted so they will match both the character itself and the standard html token. For example, the search string öl would be replaced with öl|öl .
The use of characters like åäö can slower the process dramatically and should be avoided if possible.
Finishing all the substitutions, S.O.S then makes a caseless search in all the html-files. To find all the documents, the program uses s special indexfile to keep track of the many documents located in our server. This indexfile is updated daily. However, any changes made in an existing file, will have immediate effect.
Average search takes about 15-50 sek depending on the current load of the system and the structure of the regular expression.