I have a suite of databases where I needed to achieve similar functionality, a little over 2 years ago now.
It uses a Lotusscript agent to get the parameters passed for search, start, count etc, it then does an FTSearch in the database and writes the results as a custom javascript array which it imports into a dhtml framework used to custom sort ans paginate the results.
The overhead in the agent doing the initial work is greatly enhanced by the fact that the resulting search data is cached in the browsers javascript array.
It takes just a second to return the results, and as each line is only about 200 characters 1000's of results take little transfer time, and processing on a FT index database is one of the fastest searches anyway.
The javascript code that gets included with the results then allow sorting and resorting of the data by columns, without another round trip to the server, so again is very fast.
It also allows muti-word searches, which is typically something the native dominosearch functionality generally doesn't like.
I have a suite of databases where I needed to achieve similar functionality, a little over 2 years ago now.
It uses a Lotusscript agent to get the parameters passed for search, start, count etc, it then does an FTSearch in the database and writes the results as a custom javascript array which it imports into a dhtml framework used to custom sort ans paginate the results.
The overhead in the agent doing the initial work is greatly enhanced by the fact that the resulting search data is cached in the browsers javascript array.
It takes just a second to return the results, and as each line is only about 200 characters 1000's of results take little transfer time, and processing on a FT index database is one of the fastest searches anyway.
The javascript code that gets included with the results then allow sorting and resorting of the data by columns, without another round trip to the server, so again is very fast.
It also allows muti-word searches, which is typically something the native dominosearch functionality generally doesn't like.