|
CSE Home | About Us | Search | Contact Info |
|
Anonymous feedback You can submit anonymous feedback to the instructor and TAs using the following feedback form: http://www.cs.washington.edu/htbin-post/unrestricted/education/fb.cgiIf you give permission for us to repost your feedback, we'll append it below on this page.
Date: April 11, 2011 2:55:06 AM PDT Hello, I don't usually do this, but I feel like course evals are always too late to let you know how things are going, and don't give adequate freedom to say what I'd like to say: The course content so far has been fantastic. I have been blown away. It is astounding. The lectures are well thought-out, the slides are extremely precise, organized, and informative. Your voice is clear, your follow-up questions ("why is that?") are well-placed and engaging, and your speaking is confident. The sections are very well-organized, and the content thoroughly covers relevant material from lecture. The slides are great as well. The exercises, especially the solutions, and homework 1, are extremely well-written. The quality of the code is better than anything I've ever seen; it is entirely in another league. It is crystal clear, which is the best thing (in my opinion) you can say about code. This is so refreshing, especially after classes that rely so much on recycled material. Of course you are drawing from many sources, but the compilation speaks of the time you (all) have put into this. Spot on, excellent job. I'm really looking forward to the remainder of this class. Thanks for reading. Submitter checked "permission to post publicly" box: yes
Date: April 29, 2011 1:19:10 PM PDT Hey Zork, I'm real proud of you, and I'mma let you finish, but Adventure is one of the best interactive fiction games of all time. Of all time! Submitter checked "permission to post publicly" box: yes Editor's note: Yeah, adventure is clearly the top dog in the field. For kicks, you might want to read this short history of interactive fiction; I never new that this genre was inspired by actual cavers exploring Mammoth and Flint Ridge, in an attempt to prove the caves were all interconnected!
Date: May 15, 2011 11:44:43 PM PDT So far I'm enjoying the class a lot, but there is one substantial problem with it: The feedback cycle for the homeworks is longer than the time between due dates. It's possible that I'm working my fool head off on HW3, and will find out only too late that I'm practicing bad habits. This prospect really grinds me. I appreciate that its a substantial amount of work to grade these assignments, and I appreciate that a lot of work went into designing them–I like the assignments. But feedback between due-dates should be a design constraint for them. Please, consider this in the future. Thanks! Submitter checked "permission to post publicly" box: yes
Steve's reply: I empathize, and completely agree
that it would be much more helpful if we could get the (N-1)st
assignment back to you well before the (N)th assignment is due. We're
currently operating with about 2 or 2.5 week grading latency, which is
decent but should be better. Part of this is because we're creating the
grading rubric and infrastructure as we go, and part of this is
because we're learning what it means to grade the projects, as this is
the first course offering. With any luck, what we learn this time
around will help the next course staff grade more efficiently.
An alternative would be to space the projects further apart, but
that wouldn't work well, since there are only 10 weeks in the course
and we think all four assignments are valuable. That limits us to
about a two-week cycle, since we can't reasonably hand out the first
project during the first week, and we don't want to have a project due
at the very end of the course (so you can prep for finals).
Thanks for the feedback; I appreciate it. --Steve
Date: May 18, 2011, 5:14:37 AM PDT Judging by the Catalyst Gradebook, when people do extra credit on homework, their score can exceed 100%. This would skew the mean upwards, which would harm the grades of people who do not do the extra credit. Extra credit is intended to be extra, and should not affect the grades of those not doing it in a negative way. I hope that the staff will take this into account when calculating grades. Submitter checked "permission to post publicly" box: yes Steve's reply: 100% agreed. Extra credit is exactly that, extra. What we will be doing is calculating everybody's grade as though nobody had submitted extra credit. Then, we'll be using any submitted extra credit to add to individuals' scores. So, if you don't submit extra credit, your grade will be exactly the same as if we never had any extra credit in the first place: it's still possible for you to get 100% on the class.
Date: May 28, 2011, 3:49:29PM PDT Hi! I was compiling echo_concurrent_processes.cc and forklatency.cc and both files needed to include <sys/wait.h> to get rid of the compiler error. Just thought I'd let you know :) Submitter checked "permission to post publicly" box: yes Steve's reply: whoops, good catch, yes. I've updated the web page with the include added. Interestingly, the code compiled without a problem on Mac OS X, which is why I didn't spot it. This is a common issue: not all UNIX variants are 100% identical; some will require slightly different includes, or define types slightly differently. If you're curious, check out "autoconf" -- it's a tool that can help mask these slight differences, though it's a pretty complicated tool.
Date: May 30, 2011, 9:32:18PM PDT As I'm working on HW4, I think one of the most difficult things is conceptually getting the handlers, listeners, events, threads, and everything in a clear hierarchy to know what's going on. I think the diagrams in homeworks 1 through 3 were incredibly helpful, so if possible (for future offerings of 333) I think a visual aid for homework 4 would be extremely beneficial (at least for me). Submitter checked "permission to post publicly" box: yes Steve's reply: 100% agreed again. To be honest, I just ran out of time. The event-driven, asynchronous I/O style of programming is quite non-intuitive, as the sequence of steps to process a request is now scattered across a bunch of different event handlers. Understanding the relationship between them requires tracing a bunch of registrations and callbacks, and that can be painful. I'll see if I can get a diagram out in the next day; I completely agree HW4 would benefit hugely from one. Thanks for the feedback... |
Computer Science & Engineering University of Washington Box 352350 Seattle, WA 98195-2350 (206) 543-1695 voice, (206) 543-2969 FAX [ comments to gribble ] |