I'm in Tasmania for most of this week, then off to Melbourne for a few days of fun, then back to Ireland to carry on the effort that is a PhD. I am attending the Australasian Computing Education Conference (aka ACE 2006). I'll be presenting a paper here on Thursday morning (Wednesday night to everyone else). I'll post the slides here when they are ready. The paper is about a study I ran in Maynooth analysing the effectiveness of automated assessment, and its correlation with traditional assessment techniques employed in most universities. Anyways, I have some things I'd like to write about...
Tasmania is an extremely beautiful place. The weather here is perfect, every day is like a hot summer day in Ireland. That's great, because that's about as much heat as I can handle while staying in a good humour. The landscape reminds me of a cross between the Caribbean and Scotland, in that there are loads of hills and mountains, but they all look very nice. Once I walk more than 20 minutes away from my hotel, sights like the above are reasonably common. I have taken some of my own pics, but can't be bothered connecting the camera to the laptop, and uploading them. Plus, I am not a professional photographer, nor do I have any aspirations of being one. Oh and by the way, there is a such thing as a Tasmanian devil, I saw one in the wild on my here here. Disgusting creatures, here is what they look like.
Anyways, enough about the scenery. I was at a very interesting talk this morning about "Chick Sexing" and how it relates to computer science education. Chick Sexing is something done in chicken pens, where the basic goal is to identify all the male chickens a.s.a.p as they will be first killed and let the females live on. However at an early stage of growth it is very difficult to tell if a chick is a male or a female. This is what the chick sexer gets paid for, he basically looks at chicks and looks for telltale signs that that they are male. To be trained in this job, you basically watch one of these guys at work for a couple of weeks, and slowly fine tune your ability. Accuracy for this job is in the region of 70-80%. Which means one out of every five is wrongly categorised. This method of training is called "Implicit Instruction", and it is exactly how problem solving is taught in colleges today. You sit the students down, and solve problems in front of them, hoping at some stage they'll slowly pick the skill up. If you've ever demonstrated in a programming lab, you'll know that an awful lot of students stare at a blank screen how the hell they are supposed to go about solving this problem. This is a sign that implicit instruction doesn't work particularily well.
An experiment was conducted a few years back where they got one expert chicken sexer to identify every single thing he does when identifying a chick, and write them all on a page. Then 15 random people were brought in and given a copy of the sheet, and sent off to identify chickens. These people, with no training other than a page of paper had an accuracy of 85%. This was proof to the chicken breeding world that implicit instruction is not nearly as effective as explicit. So the next time you say to a student "There is a knack to doing this", or "You'll get it eventually", bear in mind that if you actually want them to get it, explicit instruction is probably a better bet. Some people would think that its character building to leave a student suffering in front of a screen full of compiler errors, but chances are they are suffering because they are missing some explicit instructions. Telling people to read the fucking manual is well and good, but you should make sure that the manual includes explicit instruction sets as to how to solve problems. I'm sure a lot of you are thinking "Tell them to learn how to Google Des, ffs", which is a good point. What I will say, is that not everyone is born with an ability to Google effectively, so rather than rant about skills they are lacking, use the skills you have to educate them. That's why when I wrote CS211 labs, I gave students links to proper Google queries, and I gave them suggested approaches. As a result most students in the class had difficulties with some part of the assignment, and that's a big step up from a blank screen
and a blank face.
Finally, here is a picture of my dog bob.
No real reason for inclusion here, but I figured the post needed another picture. I am not strangling him despite how it looks, its just the only way to keep him still for picture, the little bastard.
Linkage
- ACSW 2006 This is the conference I am presenting at
- A sample CS211 Lab Sheet As referred to in the above text
- Arrow Code Nothing to do with this post, but a great post from a great blog about software writing.
- Open Office Templates If you want to give a presentation in the style of Steve Jobs, but don't want to pay for iWork, then look no further.