It is a truth universally acknowledged, that any scientist who wants to make a breakthrough, must be up to date on the literature. The only way to make significant progress is to be working at the cutting edge of what is known. In order to be at the cutting edge, a non-trivial amount of time must be spent reading literature. When I was just starting out as a fledgling researcher, I only had to get up to speed with the project area. My supervisors had suggested a project, and I could limit my reading and learning to within the scope of that project.

As I’ve continued in academia, both the number of projects that I am working on, and my status have increased: going from undergraduate researcher to subject matter expert and the idea generator.

At the early stages, I did not have a well-oiled routine approach to literature reading. It was a big change when I looked up strategies for how to read a paper.¹ Over time I’ve developed strategies that work for me to effectively read a paper. The next question, and the one I want to discuss in this post, is what to read. This may seem like a silly question, but given the number of papers available in today’s interconnected world - and the rate at which they are published - it is easy to be overwhelmed by the sheer quantity of papers. Without a strategy, the default becomes an ad-hoc approach which is both stressful and inefficient.

My approach has one tenet, and two stages.

The tenet: literature is a living, evolving, messy beast. I will never be able to read everything, even if I limit myself to the narrowest of areas. Therefore, I am content to be broadly speaking ‘up to date’ but will never ‘have read it all’.

Stage zero: using content aggregators/bloggers.

I could spend an untold number of hours reading everything available. However, this is not a good use of my time. The literature is a tool to know what has been done, so that I can then spend my time doing the next thing. I use a few bloggers to broaden my exposure to high quality research.²

Stage one: getting up to speed on new areas.

When I have a new area or topic that I want to get across, I do a deep dive. Initially, I check whether there is a wikipedia entry. Though not 100% reliable, wikipedia is an excellent starting point for quickly getting across an area (assuming it’s a sufficiently popular area - you know it is a specialised area when wikipedia doesn’t even have a stub). I then use a science search engine (scifinder or google scholar mostly) to search for keywords. The idea is to find the key papers, the key players, and the key concepts involved. A recently published review article can streamline the process, presenting a good starting point for getting to grips with the field. I always go back and read the main papers cited in the review.

Once I’m up to speed, I setup an alert (I use scifinder) to send me emails when a paper is published by one of the key researchers, or using a key word. This way, I can be confident that I formed a good baseline, and that I will be kept abreast of any new developments. This allows me to remain a subject matter expert without having to spend innumerable hours digging through the literature to find the gems.

Stage two: skimming some key journals.

Chemistry is a fantastically broad discipline, and I find interesting papers in all sorts of unexpected places. I subscribe to the RSS feeds for a few journals. I have two half hour sessions scheduled in my week (Monday morning and Wednesday afternoon). During this time, I skim through the recent articles. I look at the title and authors. If something seems a bit interesting, I’ll read the abstract. If I’m still interested I add the article to my Zotero library, in whichever folder best matches the paper’s area, and mark it with the tag “to_read”. In a separate session, I go through these papers in more depth. This allows me to see what research trends are emerging, as well as stumble onto completely new ideas. I find that reviews are sometimes excellent starting points for emerging concepts or areas of contention.

Actually reading the paper.

Reading a paper requires a specific mindset and strategy.¹ No matter the strategy used, if the ideas and concepts expressed in the paper remain ephemeral, stored in my head, then in a year’s time I have most likely forgotten what I read. I use a process of progressive summarisation to distill out the key takeaways and extract the most out of what I read. First, I read the paper (that has been selected previously as interesting) with the intention of taking notes. I highlight important sections using a three colour code: yellow for things relevant in the paper, green for things relevant more broadly, and red for concepts I disagree with or don’t understand. I also annotate sections with ideas and where they fit into my broader interests. When I’ve finished the paper, I change the tag from “to_read” to “to_process”.

Return to the paper and write the summary

After at least a week, I revisit any papers tagged “to_process” and re-skim the paper. I condense any highlighted parts into summaries in my own words and put them into my zettelkasten.³ I also note down any standing questions, or ideas generated from reading the paper. Once processed, I remove the tag “to_process”.

To keep the insights gleamed in a useable format, they need to be linked within my zettelkasten. The most basic form is to find the most relevant Map of Content (MOC) index and add the literature summary note onto that list. However, I also try to find other literature, concepts, or question notes that are relevant and link those.

This process builds up a map of the research space and my thinking on it. I can easily see where I have lots of understanding and where I am lacking. If a concept keeps popping up that I don’t quite understand, then I will start a note for that concept. Similar to the literature summary, I will read a textbook explanation and try to get to grips with the key points. Then I write an explanation in my own words.

Do this process regularly and in a consistent, disciplined manner

In order for literature notes to be effective, they have to both exist and exist in a useful format. To meet these needs, I have a schedule and a template for taking literature notes.

I have a regular hour long time-block on Monday afternoons for reading the interesting literature (the one’s I marked with “to_read”). Sometimes I quickly find the paper was not what I thought it was, and I decide that no notes or summary are worth my time. Otherwise, I read it intently, with the goal of later taking notes.

I aim for two hour-long blocks later in the week to convert “to_process” papers into zettelkasten worthy notes. Sometimes it takes less time than expected. Sometimes so many concepts and thoughts come up that writing and connecting all of them takes longer than expected. The writing of the summaries and accompanying notes is a very mentally draining activity. I requires deep concentration and uninterrupted time. I usually do it in the mornings when I have the most clarity, and have found that it is not something to do in a post-lunch slump.

This process of sequential summarisation and connecting concepts takes more time than just skimming papers in an ad-hoc fashion. However, the sum is greater than the whole, and, as my zettelkasten builds, my understanding in certain areas is improving beyond what I could achieve if I just tried to retain information in my head. In order to stand on the shoulders of giants, we have to understand what they have worked out so far. We need to read the literature.