We have just published an excellent guide by James Fraser on using RSS to find relevant new research papers. James does not believe that staying current on publications is really a problem that requires a service like PubChase to tackle it. His proposed solution relies on identifying about a dozen most pertinent journals and scanning new publications in them with RSS, along with keyword and citation-based searches.
Below are our responses.
——————————————————————————————————
[By Matt Davis]
Some people are both keenly observant of street signs and also have an exceptional sense of direction. When they go to a new city, they just need a quick initial glance at the city map, and then have no need for Google Maps on their iPhone to get around.
But if the goal is to keep people from being lost, then Google Maps certainly helps advance that goal.
So, one could do everything you’ve described, and if one had your reportedly well-above-average ability to read and absorb material, and the dedication to stare at his/her iPhone during each spare moment that you describe, then one would find it as easy as you do to keep up.
But most people don’t meet those prerequisites, so why not make it easier for them by using a little statistics to prioritize things for them?
——————————————————————————————————-
[By Lenny Teytelman]
Though my opinion as a co-creator or PubChase is clearly very biased, I think there are a number of strong arguments against the RSS approach as proposed by James.
1. As Matt Davis (the original force behind PubChase creation and recommendations) has written before, he wanted to build PubChase to move scientists away from relying on tables of contents (TOC) of high impact factor journals. What’s wrong with focusing on the TOC of the most popular journals? The problem is that the average PubChase user has articles from 43 different journals in her library, and 15% of our users have over 100 journals. In my small personal library, the 248 papers that I have are from 69 journals. Counting the total articles from my top dozen journals captures only 144 (58%) of the papers in my library.
2. To supplement the TOC feeds, James recommends setting up a series of author, citation, and keyword-based RSS feeds. To me, this is an onerous setup process. I have to think of all the relevant keywords, but not too many or I will be overwhelmed with matches. Which keywords? Which authors? Why do I need to spend time setting this up, if an automated algorithm can capture my interests automatically? I have not gone through the steps in Jaime’s guide to actually set up the feeds, but it does not look like a 5-minute process. In contrast, to quickly set up personalized PubChase recommendations, I just need to import my citation library as a .bib file from Endnote, Refworks, Zotero, or any other references manager. This can be done in 1-2 minutes.
3. Jaime writes, “I find it very easy to stay on top of the literature. I think the “easy” may be due to how fast of a reader Jaime is. He writes that when he reads a paper it takes him 3-15 minutes per paper. Knowing James, I am sure that this estimate is correct. However, when I read a paper, it is usually 30-180 minutes. So, if I read 10 times slower than James, then the 20-minutes of daily TOC scanning would be 3 hours for me.
4. There is also a moral argument against the tables of contents of the top journals – doing so props up the reign of the impact factor. If articles become discoverable no matter where they are published, there will be much less reason to waste months and years on trying to publish in the journals with the highest impact factor. This would be a tremendous boon for reading, publishing, and in general, being a scientist.
1. You’ll note there is a pretty wide range of impact factors in the journals I follow, so I reject the premise that it has anything to do with popularity. Additionally because of the author, keyword, and citation based feeds, I would imagine that I probably see articles from ~50 journals each month.
2. [Regarding how long it would take to set up RSS]
I went through this with people in my lab this week, and it takes about 15 minutes. It might take longer to read my original diatribe and this followup! I’m all in favour of many more sources for new papers – our lab motto is beer AND tacos, not beer OR tacos. I worry that PubChase encourages a narrower view of the literature than would be realized if the search space were broadened a bit more by unbiased journal TOC gazing and author, keyword, and citation-based feeds.
3. I am a fast reader - but I think that learning when to stop reading and to read at different levels of depth is an important skill for scientists.
4. I follow several single digit impact factor journals and I think by keyword, author, and citation searches, I actually scan a wide non-impact factor-biased section of the literature.