A fulltext indexed search is very different from a text search and you should work hard to understand the difference.
Straight (i.e., non-fulltext) searching through the text in even thousands of records is very fast to start with.
"SELECT * FROM table WHERE textcolumn LIKE '%ABC%'"
Typically you would index associated columns to reduce the number of records where text is being searched.
table
zipcode
timestamp
type
status
textcolumn
You could easily and cheaply index zipcode, timestamp, type, and status == assuming all of these are very simple integer or varchar columns etc.
So if you had a million records where you had text to search, you'd first try to reduce the search based on your indexed columns
SELECT * FROM table WHERE textcolumn LIKE '%ABC%'
and zipcode ='11111'
and timestamp <'20040202000000'
and type='dumb'
and status='resolved'
MySQL would automatically use the indexed columns to reduce the set of potential records.
Only after finding a set of records that matched the zipoced, timestamp, etc., would the text be searched. The number of text-search records is therefore likely to be very smaill.
This is why Ebay asks you what category you want to search for the term 'Rolling Stone' -- the text search is much faster if you predefine the search to be in the 'music' category.
You would want to index your records in a similar fashion to the example above even if you intend to use a fulltext search (MATCH AGAINST) rather than a text search (= or LIKE). This would reduce the number of records finally returned in a 'fulltext' search.
