csc148				Comments from the Markers
Spring 1996			of Assignment #2
===========================================================================

Read this feedback carefully, and try to use it to improve your future
assignments.


Q1.  Defining the specs:  
------------------------

     Generally well done,
     although some students' specs did not contribute at all to 
     distinguishing between the mazes in the two figures.
     In general, the simpler the spec was, the fewer problems it had.


Q2.  Maze checker program:
--------------------------

     Some people included no internal documentation (what important loops
     were doing, etc.) so at times it was impossible to know quickly
     what exactly was going on.
      
     The most annoying part was that people did not annotate their output
     so a lot of times there were up to 50 pages of output with 1 printed
     line on each page and no comments on what it was for etc.

     Many students did a lot of tests but they forgot to test the
     original cases from Q1.

     *** SUGGESTION: Use the "Printing Turing Assignment" command in the
     OOT environment.  It produces a nice package with all of your test
     runs, saves paper, and uses smaller fonts that avoid wrapping around
     lines.

     A handful of people did not discuss the limitations of their program.


Q3.  Ambiguity:
---------------

     This was in general poorly done.   Most students' answers were 
     still ambiguous.  A lot of people did not even refer to the specs 
     in their answer. 

     Suggestion: Be sure to look carefully at this part of our solution
     and ask us if you don't understand it.


Q4.  Maze builder program:
--------------------------

     Same as Q2.  Those who did the best did a thorough job of convincing
     me that their program worked.  Others did not have a complete test 
     suite and were marked accordingly.


Q5.  Data structure:
--------------------

     Well done in general.
     Some students did not understand what data structure means -- they 
     described their algorithms.


Q6.  Users' guide:
------------------

     Well done in general.  For this program, a step-by-step guide
     seems to be the most clear and effective way to guide the user.
     Few were able to clearly articulate the instructions in prose 
     Most people forgot to discuss what happens to invalid inputs.

     In many cases, the User's Guide
     was not written for the "user" which should be the appropriate style
     for this document (as opposed to a reiteration of question 1).


Quality of Writing:  
-------------------

     This could be improved.  Simply proofreading 
     would help immensely -- obviously, many people didn't proofread.  

     Some students had grammatical problems.