Asking the Right Questions

At the start of DevBootCamp, I wrote a blog post on research. But every good research starts with the design of a good question. If your question is too vague, it’s likely that you will fail miserably when doing research on that question.

Generally speaking, it’s always better to divide a broad question into narrow ones. Trying to ask people how to create a Space Invaders game will lead to failure, but asking people how to program a movable ship, how to program barriers, how to program enemies, etc. can lead to real progress. You are more likely to find sources that can easily answer those narrow questions, than you are in finding one source that can answer the much broader question.

You are also more likely to realize what you already know. It’s possible that you already know how to program movable ships and barriers, so you don’t even need to conduct research on stuff you already know. Instead, you can just focus your efforts on researching how to program enemies instead.

But what if you can’t divide a question down any further? Then you have to add some way to quantify and measure your success. How close are you to finding the right answer to your question? Now, your quantifiying and measuring does not have to be “quantitative” (using numbers). Your measurements can be just your gut feelings if you are more comfortable with that. But you do have to measure your success somehow.

Measuring success is important, because it allows you to know whether you are making progress. If you have started browsing for the answer and there is no progress being made, then you have to decide whether it makes sense to continue onward with that specific question. If it does not make sense, then you need to come up with a new question that captures the “essence” of your previous question and then start doing research on that. The answer to your new question will, more than likely, provide you with enough information to answer your original question.

You may have to create and throw away multiple questions several times to get onto the right track.

An Anecdote To Demonstrate Throwing Away A Question

Recently, I wondered whether I would have an “RSS feed” for this blog. An “RSS feed” is a simple way to check what updates were made on a blog…without actually going to the blog itself. Numerous websites have it, and even the default Jekyll-generated website post has its own RSS feed. This made me wonder whether I should also have an RSS feed too.

My first research question was: “Should I have an RSS feed?”, but then quickly realized that question is way too opinion-based. I realized this when I clicked on the first blog post and read Danny Brown (a blogger) criticizing RSS feeds. After all, the blogger claimed, most people uses email subscriptions to keep track of blogs instead of using clunky RSS feed interfaces. This may have sastified me if it was not for the commenters on that blog, many of whom use RSS feeds and disliked Danny Brown’s criticism. These users said that they find RSS feeds to be very helpful in reading multiple blogs at once, and dislike the idea of receiving hundreds of emails. Skimming the comments also reveal that there were several commenters who agreed with Danny Brown and preferred email over RSS.

I did not really want to spend time wading through the pro/con arguments though, and did not really care about the relative merits of RSS feeds versus email subscriptions. I just wanted to know whether I should have a RSS feed. But the question “Should I have an RSS feed?” suggest that I have to weigh the pros/cons and figure out what is the best argument for having/not-having an RSS feed. Success with this line of questioning would not be worth the costs.

So I instead changed my research question to “Are there any uses for RSS feeds that would justify me creating them?. Presumbly, if there are any uses that would justify me creating then, then I would create them.

Already, I can see some use for RSS feeds: attracing those people who find RSS feeds helpful. Those people may make up a minority of your readers, but they do exist, and it may be good to help them, in the same way that you would want your website to work on older web browsers. But it still takes some time to maintain the RSS feeds to ensure they work properly, and that is time that could be spent elsewhere. So I continued searching.

This blog post showed me multiple uses of RSS feeds, including auto-tweeting your blog posts and managing email subscription services. It seems that while humans may not like RSS feeds, robots can more easily extract data about a blog post from them and use that data in convincing ways. So, if I want to use these robots, I should create an RSS feed first.

In only two clicks, I have finished my research and determined that I do need an RSS feed. This is, of course, a lucky example, and one not likely to be replicated. During this blog post, I did four to five clicks (and trial and error) trying to learn how to link to my own blog posts in Jekyll.

As a useful teaching tool to explain how to properly throw away research questions, I think this anecdote works.