/home/dayanandasaraswati

I’m back after a big period of gestation. Let me begin my new season with a post on what I learnt in this gestation period - How to do research?. This is one question which most people barely dare to ask. This is because of the common misconceptions like

• Research is for esoteric eccentrics only.
• Research will make you a nerd
• Research is only for people who have a gifted cognitive power.
• Research is only for people who want to keep studying till the end of their lives. Practical people ***Do Not*** do research.

This list goes on. These are just lame excuses for not doing research. It is indeed true that not everyone of us can become Mr. Einstein. But research is not about becoming Einstein. It is just about understanding a problem and providing it a solution. Dont we understand problems in our daily lives? We all do. For example, if you find that your pants are too tight, it means that you’re becoming a glutton and you gotta be a little stringent in your diet :D. You can frame millions of examples where you understand the problem and find the solutions to it. What does this mean..??? This means that We All Do Research Daily.

Good. Now that you’re convinced that you are a researcher, lets move on to the one and only step in becoming a professional researcher. The one and only mantra is

UNDERSTAND THE PROBLEM

How is that understanding the problem completes research? Dont we need to frame the solution? Human brain is such a beautiful machine. Once you understand the problem fully, it’ll provide you with the solution without explicit efforts. Many people have their tricks of understanding the problem. I choose the following technique.

1. Decide the area of  research. This can be a very broad topic. For example lets choose “Algorithms”.
2. Now the topic is too vast. We must narrow our topic further. Since algorithms exist in every field of computer science, lets narrow down to “Operating Systems”
3. Now that our topic is “Algorithms in Operating Systems”, we now go to the abode of researchers - scholar.google.com. This is a site created by Google to provide specific search results to researchers. You will get only research works as your search results. Scholar is awesome because along with a research work, it also provides link to the site where the PDF document of that work is available.
4. Search in Google Scholar for survey papers on “Algorithms in Operating Systems”. Your search query could probably be “survey of algorithms used in operating systems”. Your search results could be varied because there are different parts of the operating systems where algorithms are used. For example, scheduling algorithms, memory allocation algorithms etc. Collect  a few of the survey papers and download them.
5. Our next step is to read through the survey paper. You’ll get a lot of idea regarding the kind of work that has been done till date. Make sure that the research paper is sufficiently new. After reading a couple of survey papers certain class of algorithms will interest you.
6. Now make your search specific to that area. Find out more survey papers in that area. Important areas to look in survey papers are the “Conclusions”, “Future Work” and “References”. The first two can give you an idea of what is going to happen in the future. They can even be a starting point for your research work. Before starting your research make sure that no body else has already done the same work. References can be critical sources of information. When reading the paper, you may not know a lot of terms. Go back to Google Scholar and download the reference papers and read them. Citeseer can help you easily track the references.
7. After the above step, you will be very close to having your own idea. Most of you would have got your idea in place. Dont narrow down on one single idea at the first place. Have a queue of ideas. Search for availability of existing implementations of those ideas. If someone has already implemented the idea, download his paper and read his future work and references. Do this for all your ideas and you’ll finally have a more refined, polished queue of ideas. Now among all these ideas, after delibration with your research guide(if any), fix up an idea and start working on it.
8. Working on the idea involves reading all related technologies to your idea. Collection of all necessary papers and reading them thoroughly. Finally concretize your idea and implement it.

As you can see from the above steps, only the last one is concerned with implementation, more precisely the solution for the problem. First seven steps bother only about “Understanding the problem”. This is the simple algorithm I’m following to do my final year project. All these are not my   own treatise. They have their origins from my friends’ advices.

I’m also following the same procedure to do my FYP. If you guys find a flaw in that procedure, please let me know. It’ll greatly help me too :P. Happy researching. Bye..:D