Yesterday I visited my adviser to go over my past exams. I wanted to go over both of them. I said that we could meet at a later date in case we did not have time during the current meeting. He said that it had been to long since my first exam and they no longer had it. This would be the only exam we could go over. I ask him how we should start. He said any way I wanted. So we started at number one. The first problem was one where I had to draw a class diagram using patterns to solve a particular problem described. Because I missed the word why in the question to cause me to explain why I chose the patterns I did I lost credit. I believe that the language I used in the description was good enough to explain the why. I talked about how one of the patterns was a creational pattern. I talked about how a creational pattern worked. In a different part of the same question I was ask about how I would analyze my design against a functional requirement to succeed in 3 seconds. I described how I would use some timed analysis to analysis my design and gave a pseudo code to gain information about the run time. He said that I was supposed to describe the different ways of analysis and not a specific one. The question was not specific as I could see in this regard. I saw that this was a subpart of the first problem and that they wanted to know if I could do any kind of formal analysis over my design. In my interpretation of the first question I passed this question. It was poorly written and did not attempt to draw anything out of the student. It relies on the trickery of well placed interrogative adjectives to cause the student to stumble on a timed test with stress. This was a trivial question but made difficult by the manner in which the question was asked. This is s central theme of the exam. The next question was on different types of testing. I am personally confused by the wording of the names of types of testing so this kind of recall I get wrong some times. There were three of them. I thought for sure that I had at least two of them. I am pretty sure that I did but because I described how a test would be carried out rather than the selection of testing questions in that technique I did not get the question correct. Again, this sound like a poorly stated question. I think I had at list 2 kinds of the testing correct. I just talked about them in the wrong way per what the test writers wanted....subjectivity. The third question that I selected to do was on formal methods. I don't remember why I chose this question. The question was on the language OCL. OCL is used in UML diagram notes boxed to describe pre and post conditions as well as parameters. OCL is actually a language though in real life it is never used that way. I new that there was more to it but had decided that this has a small chance of a significant part of the test and didn't go deep enough in my study. I had a handle on the syntax but was unable to put it together on the test. My answers ended up making no sense. I wouldn't say that I was completely wrong but certainly you could ague it that way. The last questions were questions like "What is Deadlock?" I had this very same question on the Operating Systems exam of which I passed. I wrote the definition directly from the book. In fact for every question in this section I wrote the definition directly from the book. I was not given full credit because they said I had not demonstrated that I fully understood what each question was about. My adviser and I argued about these questions for some time. I argued that a definition should be sufficient. He argued that a PhD is not about amassing knowledge....his words not mine. He wanted pseudo code to explain my answers further. I said that that does not demonstrate anything. If a person memorized the definition he or she could memorize pseudo code also. This is a slippery slop to define the correct answer by something that you could not memorize. He said that the large body of knowledge needed for the test would prohibit that. I don't think that a large body of knowledge prohibits anything. Sounds like the Turing test to me. I think that I should be allowed to continue on the basis of the last test, in my mind I passed. I know that when I am in a stressful situation like a timed test there are certain things like definitions that get me through. If they want me to elaborate on every question a timed test is not the way to test me. I also find a definition somehow definitive. To explain more puts me into a thrash that is hard to dig myself out. I continue to want to go back to a question and never finish. I find myself being repetitive and making the answer hard to read. On a test where there is one place to put your thoughts down on paper a thrashing situation means that organization is out the window because my organization in my head is nonlinear. My organization is more like a 3d web. To draw knowledge and link to the place where the true answer is means a depth or breadth search of disjointed data translated into linear answers which often times don't make sense because of the peppering of irrelevant digressions. I am really tired of tring to explain. Having an untimed test may help but if the test is completely subjective how can I ever get a fair outcome. It seems that the subjectivity has to be removed. Answers need to be concrete and above reproach. This technique didn't seem to work in my last test. I thought that a definition would be sufficient. I see now that if the test is subjective even a definition can be said to be wrong on grounds that memorization of a definition does not demonstrate understanding. I remember reading about Gauss and how he often had short terse answers. His notebook, published after his death, was only 19 pages long. Would he have passed one of these tests? Does short terse answers not constitute understanding? Why does an answer of a definition, yes obtained from a book, not be the answer? It seems that the technique of denying a student the attribute of understanding, the burden of proof should lie in the test and not on the student. Should a student make the assumption that every grader has no prior knowledge and teach them about their answer. It seems that the audience for any test is the tester. If the tester is inadequate in their understanding of the answer and the answer is correct they are then a researcher learning themselves. So, the student does become the master. If the tester with their limited understanding, but in the position of authority, cannot adequately grade the test then they should be removed or the test be made so that anyone could grade the test with the correct key. Thus, subjectivity should be the last tool lest they believe that they are smarter than all who may take the test. Objectivity is the only true test, not just because of the lack of influence of personal feelings or opinions but because of the possible lack of understanding on the part of the tester. This lack of understanding could be either in the answer or in the test and what the test is supposed to show. I believe that the test is flawed. I don't believe that the professors (tester) are so arrogant that they believe that they are smarter than all who take their tests. My last point that I will make is that I have taken 3 candidacy exams. Each given on a different subject: Operating Systems, Database, and now Software Engineering. There is not any consistency among how the tests are written and how they are graded. Some have concrete questions some questions some don't. Some have essay some don't. I have gone to see the exams in the past for other tests and they way they are handled is differently. Some have the test marked up and some don't. The software engineering test had no marks on it. My adviser had to interpret what was wrong with the test. I understand that it is graded by multiple people. If this is the case, lets see what they all have to say. Lets not rely on what my adviser has to say about each answer. This process of reviewing the software engineering exam could have been done better. The student goes to get understanding. I think that I could have gotten more out of it if I had seen all of the other graders remarks. After all, I am trying to please each and every one of them.