Sunday, October 31, 2010

scheduling issues

Due to unforseen events unrelated to this class, I was unable to attend the review session; further, I won't have office hours today. Sorry for the sudden change. so, I will have office hours tomorrow, Nov 1, 5:30-6:30pm. So sorry for the confusion. This was unavoidable today.

Friday, October 29, 2010

2nd Midterm Review Posted

I just posted the second midterm review on the lectures portion of the portal. Good luck! As a reminder, the review is this Sunday, from 2-4 PM in WEH 7500.

make-up office hours

Since I missed yesterday's office hour, I will hold one this afternoon from 2-3pm.

Thursday, October 28, 2010

Late for office hour

I will likely be late for my office hour by about 1/2 an hour, possibly more. Sorry for the inconvenience.

Wednesday, October 27, 2010

Graphs Lab Released

Graphs lab has been released on FrontDesk. For this assignment, you will be allowed to work in pairs for the programming portion, however, the theory is still an individual assignment. The theory is due on November 3rd, in recitation. The programming portion of the assignment is due on November 11th. This lab is very long so we strongly advice that you start early!

Monday, October 25, 2010

Update to seamcarving assignment

I just updated the seamcarving assignment on FD to address an ambiguity that was pointed out with multiple valid lists. An example of such a list is {1,9,9,1} for which both {1,9} and {9,1} are valid minimum weight subsequences. Note, however, that an example like {1,3,2} only has one valid min-weight subsequence, {1,2}, since it was specified that your list-finding algorithm should find the longest (greediest) valid subsequence. With respect to logistics, any new submissions made to FD support this change. If you've finished the assignment, I'd highly recommend resubmitting to make sure that your code still works (it should). Furthermore, if you download a new copy of the stub code, the runTestTuples() method you get has been adjusted to reflect this change. Good luck!

Tuesday, October 19, 2010

seamcarving clarification (theory)

For 2(f)(i), you can just let T(n) be the runtime of Stack(n) (not the solution to finding the tallest stack). If you do it the other way, you'll still receive credit. For 2(f)(ii), 2(f)(iii), you should actually try and find the tallest stack, not just compute Stack(n). Sorry for the confusion.

Monday, October 18, 2010

Makeup Office Hours for jpaulson

I missed my office hours this evening; I sincerely apologize to anyone who went looking for help and couldn't find me.
To atone, I will be holding makeup office hours tomorrow evening from 6:30-8:00pm in the Wean 8 couches.

Seamcarving Clarification (Code)

Your list-finding algorithm should be greedy; that is, include as many elements as possible. This is a generalization about the comment made with respect to 0's in the documentation (that they should be included). If the list is [3,5,2], you should return [3,2]. Similarly, if the list were [0,3,5,2,0], you should return [0,3,2,0].

Sunday, October 17, 2010

Office Hours

Due to me not being in the appropriate state, I won't have office hours today.

Friday, October 15, 2010

Typo in seamcarving theory

There was a typo in the seamcarving theory assignment. In 2(e) it should be "max" instead of "min". It has been updated on the lectures portion of the site.

seamcarving released

The seamcarving lab has been released on FD. Please be sure to complete the theory first, because its solutions will be invaluable for the programming component. The theory has also been (very slightly) updated, to reflect a new number of total points on the lab (138). All of the questions and content is still the same.

Wednesday, October 13, 2010

Anonymous Posting on the Discussion Board

From now on, the staff will stop responding to posts on the discussion board that are anonymous. If we see such a post, we'll post a reminder of this policy and then delete the post. Please repost your question correctly identifying yourself.

seamcarving theory released

HW4 (seamcarving) theory has been released on the "lectures" section of the portal.

Tuesday, October 12, 2010

Revised slides

There is a more elegant solution to handling the problem of testing bipartiteness in directed graphs than the one I presented in class. So, I have updated the slides on the portal, and I will go over the new solution in class on Thursday.

Thursday, October 07, 2010

Compression Code Addendum

I'm encouraged to see that many of you have been making submissions to FrontDesk for the compression assignment. Keep up the good work! Below are a few comments I wanted to share with the entire class concerning this assignment:

1) You are not implementing MyPriorityQueue for kicks and giggles. You are expected to call your own code (and not Java's PriorityQueue) when you implement Huffman encoding.

2) A significant portion of the grade for this assignment is in hidden tests. The staff tests visible on FrontDesk only test very simple cases. Passing the visible staff tests does not ensure you will do well on this assignment. You need to write extensive unit tests to ensure your compression algorithms are working properly.

3) The new FrontDesk has been performing quite well and we haven't had any serious issues with it being overloaded since the switch. However, Compression is significantly more computationally intensive than Snakes, so do continue to submit well before the deadline. In addition, if you submit a unit test that takes too long to complete, your process will time-out and all the remaining unit tests will fail to run. We strongly recommend including a timeout of 1 second for all of your unit tests. For example, the following code contains an infinite loop, but will stop execution after one second:

    @Test(timeout=1000) public void infinity() {
       while(true);
    }



The advantage is that if a unit test times out, the remaining unit tests will still be run. You will likely want to run some of your more intensive stress tests locally without a timeout. If these tests cannot complete within the one second timeout on FrontDesk, that is ok as long as you stipulate this in a comment.

4) To complete the assignment you may find it necessary to perform some fairly low-level byte manipulation. This is a useful skill to learn, but it is not fundamentally what we are testing in this assignment. If you get stuck with bit mangling or byte conversions, please ask for help.

As always, if you run into problems or ambiguities, please share them with us on the discussion board!

Wednesday, October 06, 2010

2nd midterm

We decided to move the 2nd midterm to November 2nd. The schedule has been updated.

Tuesday, October 05, 2010

Quiz in Recitation Tomorrow

Just a quick heads up that there will be a quiz in recitation tomorrow. Don't miss it!