School: weeks upon weeks

School: Research Methods and Communication

This past week I finally got to submit my paper review and received two to evaluate.  I quickly realised that requirements can be interpreted wildly differently.  I had probably spent more time than had been necessary on it.  That said, I got to review a paper that was probably stressful for the author and ultimately warranted an evaluation of 0, and I'm a generous person.  The main issue was probably their understanding of what was required.  I laboured over how to deal with it, but just sent in my evaluation of it.  Yay!

I also got to give my presentation, which was delayed once by illness and once by time.  Sadly, a brief power outage sabotaged my alarm clock.  Miraculously, I still awoke with 20 minutes to just run to campus.  Sadly, this deprived me of time to setup my laptop with the projector before hand, and it didn't instantly work.  Thankfully, I had saved a slightly earlier version in PDF.  Sadly, this lacked page numbers and deprived me of my timer.  This resulted in failing two important requirements for the presentation which I must now account a dismal failure.  Ah well :) 

School: Neural Networks

I now have this large assignment to finish.  I didn't have too much time for it this past week because of ...

School: CIS*2750 TAing

So, the grades were less than stellar.  I hoped it was my automarker, but mostly it seemed to be a lack of attention to detail.  Submissions captured the broad strokes of the assignment, but in this case, nearly there isn't really good enough.  The fundamental pieces need to work well.  After that, you can afford to lose marks here and there.  Without the foundations, nothing works.

I spent a lot of time in the last week answering questions on the course forum, helping people understand where they went awry.  I went to the lab for a couple of pro-bono hours to help people understand particularly stiff problems.  One lad seemed to be trying to get me to debug all his memory issues for him, so I had to cut that short.  I would like to think that it's a lighter first year that makes this course so difficult for students these days. 

I had a LOT of fun planning changes for assignment 2's automarker.  There is a design deficiency in the largely great framework we have so far: each test in a suite shares the same process, so if one test sufficiently corrupts memory, the rest are impacted.   I've planned a change in how we call tests, forking them and killing them if they time out.  Also, fewer macros, more functions.  I've blamed my use of macros for A1's automarker on tradition and existing convention for the automarker, but I now acknowledge that the previous one used them more sparingly, and that my abuse of them was the gradual result of macro-creep and my own inattention.  Oops.

Research and NLP

So, I've a project that lingers on.  Now I have insufficient time to allocate to it, but I don't want to give up on it yet, as I still see bright avenues for improvement.  Sadly, I haven't had time for my proper research either, and have actually missed a couple meetings with my professor.  Even when I'm super-responsible in another domain, another one can suffer before I realise it.  Sigh.