Addenda and Clarifications for Project 2

Q: Do I need to call srand to seed the random number generator?
A: No - it's done in the call to InitializeExperiment

Q: Speaking of InitializeExperiment, what should its paramteter be?
A: The parameter should be N, the number of items to be used.

Q: And when should I call InitializeExperiment?
A: Just once, before doing any of the 5 experiments called for in
   the project.

Q: before we leave the topic of random numbers, what should I be 
   doing about a randomly generated fraction?
A: check the man pages. You will notice that there's a library
   function which exactly fits the bill.

Q: What exactly am I suppoesd to be counting in the experiments?
A: Each time your code obtains a value (by calling either of 
   GetRecencyDatum or GetRandomDatum) it needs to count the 
   number of comparisons made to decide what to do with that
   datum - "what to do" means - whether to insert it at front, 
   move it to front (in case of MRAF), or to insert it at 
   end (of the IAB list). You do not need to count the number
   of comparisons made to delete an item, and you do not need
   to count the number of comarisons to insert the item (which
   should be O(1), irrespective of which list the item is being
   inserted into). That is the "total accesses" data.

   Your code will also count the number of items that came from
   GetRandomDatum and GetRecencyDatum.
   Unique items is pretty self-explanatory.

   Average number of accesses = total accesses/# items.

   The sample data was not done with benefit of the program having
   been complete, so the data is off: this is what the last line
   should have been about:
				MRAF		IAE
	Average Number		1.07		5.18
	of accesses

   Average number of accesses is per "trial". 
   

Q: What about output? What should I be doing about K?
A: Each K data items your code obtains (each K "trials"), your
   code should be dumping the statistics. You should be reporting
   cumulative statistics (e.g. if K were 500, you should produce 
   output for the first 500 trials, the first 1000 trials, the 
   first 1500 trials, etc.) Then, when the experiment ends 
   (you have handled N trials), you should print out the 
   grand totals. NOTE: it is possible that N is an exact multiple 
   of K, in which case the last running total will also be the 
   Grand Totals for the experiment. In that case, you might be
   repeating outputing the same data. 

Q: It looks like the high recency data isnt too high recency. What should
   I do?
A: If you have listed  P2DataGen.o as a dependency in your makefile, just
   remake. You will get data exhibiting a very high degree of recency.
   The old P2DataGen.o is still in the project 2 directory with the
   rather unimaginative name, OldP2DataGen.o (if you want to experiment with
   it, too).
   So you should ReLink.