Problem: Computer system software must sometimes be designed to run as quickly as possible. This is especially true of large, multitasking and multi-user system like Windows NT. Your challenge is to create an optimized disk drive scheduling routine much like the ones used by advanced operating systems. The program will be a simulation of a disk driver and a hard disk.

Hard disk drives require much more time to read than to write; the hard drive your program will control requires one millisecond to read a megabyte and three milliseconds to write a megabyte. It has a capacity of 1000 megabytes, each designated with a location from 1 to 1000.

Your driver must issue read operations before writes unless a particular write has waited for 500 milliseconds or more, when it must be run before running any more reads. To make matters more complex, when there is a read that overlaps a pending write operation, the read must wait for the write to finish (this is only true if the read was issued after the write). The only time the read does not have to wait for the write is when a single pending write covers the whole read, in which case the transfer takes 0 milliseconds because it is done entirely within memory.

Input: The program must accept commands, each composed of a time delay (in milliseconds from 1 to 100), a command (W = Write, R = Read, D = Done), a location (from 1 to 1000), and a length (from 1 to 1000). It must continue to accept until the Done command is issued.

Output: After accepting the command, the program must summarize the command that has just been queued and display the operations completed since the last command. The time that has elapsed since the last command is the value of the time delay. See sample for clarifications. The program should print the current time before accepting each command.

Sample: