I hear a lot of people (not particularly in research but in other less academic areas) saying that they don’t consider reading papers past 2008 important or useful because they’re are too dated. I can understand why this perception is knocking around, but I think that ignoring older papers is a big mistake. The old papers are where you start. They are the source of everything that comes next. Without paying any attention to them it is not possible to fully understand the recent work being done in that area. Like with everything, you start at the beginning, and learn to crawl before you can walk.
This post is not only for students in computing, but also for the SEO who wants to delve deeper. I hope professional researchers will find some of the links interesting!
I’m often asked things along the lines of “How do you manage to read so many papers every week for your blog and your own research?” I learnt how to!
Keep track of it all:
The first thing I would advise is to print them out. Reading online is not the same, as we tend to scan read more and miss important details. Take notes as you go along. printing the papers mean you can write on them, and highlight things.
Store them properly using Bibtex and make very clear and complete notes on paper, in a notebook, and scan the pages in for back-up. The reason for doing this is that we tend to think more when actually writing, and work at a slower rate because we’re taking the time to think more and not mechanically type out loads of…what ends up being waffle in my own experience. Write down page numbers, the full reference, the title, the author. This very very handy when you come to writing up up either a blog post, article, thesis or paper.
Old papers are yummy:
Do not discard “old” papers. they are the foundations on which everything else is built. They will tell you how everything functions and you will be able to understand why the method or system has evolved like it has. Or why it hasn’t. This Stanfords’ Computer science reading list for information retrieval. You will notice papers dating back to the late 70′s. Paul Mcnamee from John Hopkins University also provides a reading list for his IR students and you will also see some of the same references, and they are mostly not recent. This is because it’s so very important to start at the beginning.
Mark Lee from the university of Birmingham has a nice list of popular papers that are on computer science course reading lists.
Did it take time to get through all those books and papers full of past research?
Ooooh yes! I spent the first couple of years of my PhD just reading and reading, and also most of Masters too. This is really not unusual either. You get good getting through them. I remember it used to take all afternoon to read one. When you know the basics and what all the long words mean, you don’t have to keep looking them all up so it goes much much faster!
There is no way around really. If you want to understand “NL Understanding with a Grammar of Constructions” by Zadronzny, Szummer and Jarecki for example, you are going to have to know mall about “Construction Grammar” (starting in late 80′s), which means that you have to know all about the other grammars (dating back to the 70′s), and all about natural language processing (dating back to 1945ish) at least.
If you don’t have time to read everything and you’re not specifically interested in the detailed research:
you still need to know some of the basics to understand how search engines work. Mark Curzon from Queen Mary, University of London has made available a number of PDF’s covering a whole lot of stuff that you’ll find useful for the topic of information retrieval.
I also make as much as I can availabe here in as simple a format as is possible.
Resources to make things much easier:
Michael Hanson wrote a useful paper called “Efficient reading of papers in science and technology“:
“Taking notes will help you to understand what you read and will save you effort in the future. When you have just read a paper, you may understand it well. The definitions are clear, the charts show correlations at a glance. But next week, when you are writing a report on this subject, or next year, when you need to refer to the paper again, it may not be so clear.”
Amanda Stent wrote an excellent paper called “How to read a computer science research paper“:
This is probably the very first step to take. When you are going to read a huge amount of papers, it’s good to know what to look for straight away to avoid reading rubbish or papers that end up not being of any use to you (I wasted a lot of time doing this when I started out).
She tells you where to look for them, why there are different kinds of paper, how to tell they’re authoritative, how to read them efficiently, and how to keep track of everything you’ve read.
For researches and those wanting to get into research:
There is an excellent paper to read by Richard Hamming called “You and your research“:
“One of the characteristics of successful scientists is having courage. Once you get your courage up and believe that you can do important problems, then you can. If you think you can’t, almost surely you are not going to. Courage is one of the things that Shannon had supremely.”
“I subscribe to Pasteur’s “Luck favors the prepared mind.””
And to finish with:
This actually quite a funny glossary of what things mean in all those papers. Honestly…you’ll laugh!