Li Lin @ SERL.Clarkson

Beyond Program Understanding: A Look at Programming Expertise in Industry

lildhou — Wed, 31 Mar 2010 15:18:42 +0000

This study aimed at characterizing the actual programming behaviors both retooling appenticeship and experts.They report observed strategies, skills and tools experts used in programming comparing to appenticeship. In the conclusion part, authors advocate that experts ascribe their effectiveness not only to their knowledge sufficiency and programming skills efficiency but also the principles, design rationale, experience in usage of human resources and ancillary tools.

Generally speaking, we could conclude experts as more “experienced”. Understanding towards expertise would suggest strategies to reduce effort to required to retool and relieve barriers to developing expert performance. These strategies are:

knowledge of generic programming concepts and strategies; relevant programming skills;
principles, design rationales, history and quirks of the software framework they work within;
proficiency in the use of broad range of ancillary tools;
knowledge of and appropriate use of human and written information resources;

What Can Programmers Tell Us About Framework

lildhou — Fri, 26 Mar 2010 15:16:37 +0000

This paper is generally the guidance of our current work and also includes the primitive stage of analyzing newsgroup threads. Authors analyzed JButton and JTree newsgroup threads. Their contribution lie at classify the two Swing components into a smaller scale that lead us to think of root cause. They also mentioned documentation and people issue, which inspires us thinking of programmers’ learning difficulties.

6 learning barriers

lildhou — Thu, 25 Mar 2010 23:03:09 +0000

In Ko and Meyer’s version of new identified six-barriers learning system, they absorbed previous learning barriers idea and improved the definitions so that they are now more integral. Their study aim at dividing learning barriers types and find out solution for them. In that case, they carried out an experiment that programmers learn use a language. They give some brief definition to these barriers but somehow not clear enough for further study. We referenced on this definition system and adapted it to Swing software system.

Structuring Effective Worked Examples

lildhou — Thu, 25 Mar 2010 19:02:17 +0000

Under some conditions, substituting worked examples for problems or exercises enhances learning and subsequent problem solving. Under other conditions, worked examples are no more effective, and possibly less effective, than solving problems. Based on past experience, worked examples can enhance education efficiency in directing attention appropriately and reducing cognitive load.

In this paper, authors reviewed these lab-environment based results and pointed out conditions that should be obeyed in order to take fully advantage of worked examples. They also expanded traditional areas that are cited as proof. In this paper, they chose geometric and physics for experiment material.

Experiment 1: Compared worked examples with conventional problems using lens and mirror ray diagram problems. The result shows the overwhelming advantage of worked examples to conventional teaching.

Experiment 2: An ancillary experiment to Exp 1. Claiming that both similar and transfer problems are efficient under the effect of worked examples.

Experiment 3: Tested the hypothesis that if a worked example required students to split their attention between multiple sources of information that had to be integrated mentally, the cognitive load generated was sufficiently high to eliminate any advantage of worked examples over conventional problem.

Experiment 4: Eliminate the effect of split attention by incorporating equations into problem statement.

Experiment 5: Additional, seemingly information can have negative consequences.

Above five experiment generally discussed the effectiveness and its working condition. It is not as commonly assumed that worked examples will always help inspire better than conventional education. Only under the circumstance that worked examples are not supported with redundant information and not split students’ attention, could heavy use of worked examples facilitate problems. Although this experiment is only expanded within the boundary of physics, we could still hypothesize the situation of that in software system study.

Based on our empirical study, we observed a Swing tutorial dispersion problem. Swing knowledges usually require programmers split their attention into several aspects. Even if they only need a simple user interface implemented by Swing, they still have to take care of the “global” knowledge that affects all Swing components. Therefore, if a programmer finally take all consideration of these macro-control factors, we could say they already understand Swing’s mechanism. They could feel everything under control. Now the question is: Is it necessary for all programmers to come to this stage? Is there a workaround that they could only focus on their needs and neglect other excess burdens? This study gives a conditional circumstance, that is, worked examples with a elimination mechanism of distraction and without redundant explanation could help improve study efficiency.

Enlightenment:

Epx 1& 2: In future case study, we could compare effectiveness of worked examples in different teaching phases. This study, if successful, will benefit software teaching.

Exp 3: In this experiment, test result shows us that distraction could eliminate guidance of worked examples. Analogously, I also mentioned in my thesis that Swing sample scenarios are not simple enough for end-users. It manage to include all features but seems too complex. So it is with Rob Camickr’s FAQ’s list, which is why we need to compile a new list.

Exp 4: This experiment shows how to eliminate the effect of split attention. They incorporate different kinds of materials together, build mappings between them. In Swing, if we have such kind of mappings between knowledge and practical use, design barriers will be surmounted much easier.

Exp 5: This experiment tells us that repeated explanation could have negative effect. According to me, whether the hypothesis of the last experiment could reuse on our study remains unknown.

Out of scope

lildhou — Wed, 10 Feb 2010 20:39:48 +0000

http://lilisworking.wordpress.com/2009/12/14/basic-button-feature/ OP didn’t know setText(). After searching API classes, he found a method setName() and guessed it would probably be the one needed. This is caused by confusing API.

http://lilisworking.wordpress.com/2009/10/22/pilot-study-of-jbutton-13/ Object-oriented problem

http://lilisworking.wordpress.com/2009/09/07/pilot-study-of-jbutton-button-enabling/ OOP: passing references between frames.