IT Science Research

[Snarpy]

I am taking a class called Foundations of Information Technology Science. We are covering scientific research methods and how they apply to various computing disciplines, especially Information Technology. The course will culminate in the writing of a research proposal for a scientific study of something in the IT field.

I know you all are practicing professionals and not academics and have a very pragmatic approach to the field, but I was wondering if any of you have a burning question that would lend itself to scientific investigation.

If nothing else comes along I will probably propose a website usability study, but I thought I'd get the input of people with real experience and varied perspectives.

[hobbes]

Something that could be interesting is investigating the potential applications of peer-review within IT. This is a method often used by the research community, typically involving publication and/or conferences, although it has IT (app|im)plications as well (e.g., code review).

Its not easy being #00FF00

Although it's better than feeling #00f

[kitchin]

Will DKIM fix the internet? It's interesting historically, but maybe too big a topic for your project. The neat thing about the history is that computer privacy and open source nuts fifteen years ago were braying that everybody should send their mail encrypted, or at least digitally signed. They would close all their usenet posts with their public PGP key or fingerprint. The idea was that everyone would publish a public key, which was available at a key exchange, and signed by someone else, creating a chain of trust. Then whenever you sent an email to somebody, you would sign it with your key (showing the message had not been changed in transit), and optionally encrypt it with their key (making the contents secret). Nobody did it but nerds in tights. Even though everybody knew email was insecure even back then, it just made you yourself seem insecure, I suppose. Now fifteen years later, digital signing might stop spam and fix the internet. Funny thing, eh?

[Andilinks]

I think you should choose the question that would most further your career goals and most easily translate into lucrative future assignments. If website usability is your point of departure then tilting it toward ecommerce applications and tossing in the effect of SEO would be the best pick.

A more interesting question would be about the effect of mobile web use or the effect of expanding computer power in say, a decade from now.

[Wassercrats]

My first choice would be to study why a science teacher would think that someone who takes a Foundations of Information Technology Science class would have any knowledge or interest in what needs to be studied in IT. Now some students will have to do blind research just to figure out what to propose. It's not fair to the students. And I guess every student needs his own idea. Even an expert would have trouble thinking of that many IT topics worthy of study.

My second choice would be to find some software developers who wrote their own documentation and compare aspects of their coding with aspects of their writing. That could be really good if it's done well or really dumb if it's not.

[Snarpy]

Hi guys,

These are some interesting ideas. I especially am interested in Hobbes' suggestion.

Originally Posted by hobbes

Something that could be interesting is investigating the potential applications of peer-review within IT. This is a method often used by the research community, typically involving publication and/or conferences, although it has IT (app|im)plications as well (e.g., code review).

Let me do a little mulling things over out loud.....

We've been studying what you might call the "production of knowledge" within scientific disciplines. Peer review, of course, is an important link in that chain. Applying that to the production of IT artifacts (code or whatever) is intriguing.

One of the methodologies that we've learned about is action research. Action research is a method of systematically investigating the success of problem-solving in a practical, real-world setting. A plan of action is developed to solve a problem and reach a predetermined outcome, and a reflective assessment is made as to the success of the action. The goal of action research is not to reveal existing relationships and interactions, but to trace and study changes as the result of actions. The process is usually iterative; actions are modified based on the results of the assessments. If the results of the action research are published, they are then subject to peer review - so this would be one (formal) way of peer-reviewing IT work. (Remember that action research is applied to a real-world problem solving situation.)

In one of the articles we have read (and we are reading lots of peer reviewed literature), the authors talk about using an an evolutionary learning approach at NASA Goddard Software Engineering Laboratory from 1976-2002. It seems that some sort of peer review (within NASA) must have been involved in this.

A company using proprietary code will want to keep any peer review within their own institution, but open source could solicit review from a larger community.Don't (some) open source software projects employ some sort of vetting of each other's code?

Hobbes, perhaps you could elaborate a little more on what you have in mind?

Originally Posted by Wassercrats

My first choice would be to study why a science teacher would think that someone who takes a Foundations of Information Technology Science class would have any knowledge or interest in what needs to be studied in IT. Now some students will have to do blind research just to figure out what to propose. It's not fair to the students. And I guess every student needs his own idea.

We are reading peer reviewed literature - that gives a great perspective on what needs to be studied in IT.

Even an expert would have trouble thinking of that many IT topics worthy of study.

You don't know our teacher. He's got enough ideas for all of us. I just want to get my own.

[hobbes]

Snarpy - Given the train of thought you're following above, some quick reading up on the following may help you solidify the idea (possibly in how what you end up doing differs from them):

- collaborative software development
- rapid (software) prototyping
- iterative development
- agile development

A company using proprietary code will want to keep any peer review within their own institution, but open source could solicit review from a larger community.

This is too broad a statement. I have done code review for companies/agencies other than my own. You do however tend to have a much more closed community when it comes to proprietary vs. open source.

Also note that often, in lieu of code review, developers (and/or QA) will use unit testing, falling back on code review when testing fails.

BTW, don't necessarily limit yourself to code review, as architecture/design review could also go a long way towards more robust software.