Start 4 threads which do the following task:

1. Opens every nth file (first thread will open 0th 4th 8th 12th file second thread will open 1st 5th 9th and so on)
2. Scans the file for occurrences of word 'camera' and notes down line numbers (possibly in a hash table -- the key would be the file name with complete path and value will be an array / arraylist containing line numbers)

The program would be as follows:

1. start threads scanning every nth file -- 0-4-8-12; 1-5-9-13 etc.
2. Scan the file for word 'camera' and note down line numbers in a hash table -- the key would be file path and values would be the line numbers.

but why nth file ? Why not simply divide 100 000 into four groups and have each thread run on a group ..