While 90% of my page is in php, i dont want to have search.cgi anymore. For this i make my own search now. To repeat if i have understood the indexed search right (and maybe for somebody who wants to now) i found out:

Links_Word_List stores
ID (auto_inc)
Word (the word itself)
Count (how often this word is found)

and

Links_Score_List
Word_Id (=ID from Word_List)
Item_ID (=LinkID)
Score
Wordposition

The score could be, i will test it in a minute like following:

Title, value 2
Descr, value 1

So if the word "test" is in Title and Descr it has a score of 3, when only in Descr a score of 1, when only in Title a score of 2.

Seems easy and i am glad about this, while i could easily make my sql-queries in php now.

The reason for an existing position i dont know, maybe because of high-lighting and because it is faster to count the words position than to have another regex to find the whole word?