Thursday, October 23, 2014

Trajectories in Open Access

Scientific research is a social endeavor and throughout the world, it is funded primarily by public money. Therefore, there should be no barrier to the knowledge developed through such research and it should be accessible to everyone. The primary sources of such knowledge are the articles published in scientific journals. So, to spread knowledge, these articles should be freely available to everyone. The Open Access movement, which is  spreading through the academic world, is preaching this philosophy.

In the current model of publication, researchers submit their articles to journals and the journals publish the selected few after peer-review. In this process, the author voluntarily transfers the copyright of the article to the publisher. The publisher does not pay the author. But the reader, whether a researcher or a layman, has to subscribe the journal to read it. That's what most of our libraries spend money on. Over the years, the subscription fees have ballooned to such an extent that even libraries in the developed world are falling short of their budget.  

The Open Access movement strikes at this very issue. It promotes two models, to achieve open access to published work. In the first, any body is free to access the journal over the Web. Such journals are called open access journals. For publication in such a journal, authors have to pay a article-processing fee. The cost for editorial manpower, formatting, typesetting and server management are covered by that fee collected from the authors. Over the years, number of open access journals has increased exponentially, with many having doubtful reputation. Even then, several open access journals are well respected for consistently publishing high quality work. Though such journals are promoting open science, article processing charges are often very high and prohibitive for researchers working in developing countries.

The other model, for open access, is creation of public funded open access digital archives for scientific papers. PubMed Central, developed by  U.S. National Institutes of Health's National Library of Medicine, is one such digital archive, where publishers or the authors voluntarily submit a copy of their articles. In fact, such submission has been made mandatory for every work funded by NIH of USA. Some other funding agencies are also promoting the same model. Anyone having an access to the Web can read all the papers stored in such archives. Such archiving does not violate the copyrights of most of the publishers, as authors submit only their copy of the final peer-reviewed draft without having any editing and formatting by the publisher. ResearchGate, a social networking site for scientists, also follows a similar model for sharing scientific articles.

Very recently, DBT and DST, funding bodies for scientific research in India, has released the second draft on their open access policy. They have proposed establishment of open access digital archives, in different institutions as well centrally. Any publication coming out of a work funded by these agencies must be deposited in such archives. Like other such archives, authors will deposit only their own copy of the final peer-reviewed draft. Interestingly, the proposed policy explicitly discourages author-paying model of open access journals and has made it clear that they would not provide financial support for article-processing fees. This makes sense, as article-processing fees of many journals are exuberantly high for most Indian labs. Additionally, this will also discourage spread of predatory journals, many of which are published from India.

It will be interesting to follow how India's open access policy shapes eventually. But for the time being, let me imagine the evolution of the publishing industry in the age of open access. The open access digital archiving makes sense for every country and most of the major players in science would eventually move to this model. But that may trigger a trouble. A publisher takes care of peer-review, editing, formatting and publishing. They charge the fee to the reader or the author to cover the expenses for this process. The final published articles are usually smartly edited and eye candies to readers. The authors copy of the final draft deposited in open access archives are not adequately formatted but contains all the scientific contents. Therefore, though reading such a draft is bit cumbersome, but that does not affect the science. In fact, every scientist is well trained to read such drafts. If we get such final draft without any cost, why should we pay to read a well-formatted copy of the same, published by the journal? Obviously not. Eventually this will reduce the number of subscribers to such journals. So journals running on subscription based model will not survive in the world of open access. In fact all major publishers for science journal are offering some form of open access for their journals and testing the water. But we need publishers to manage the whole process of publishing scientific articles through peer-review. At least some one has to run the peer-review process and that also have a cost though the reviewers do not get paid. If the reader does not pay, the cost has to be covered by the author. And that brings us to author-paying model of current open access journals. Therefore, institutionalized discouragement to such journals may not be a good idea as we do not have an alternate model that will sustain in long run.

There is another model of science publishing. It involves post-publication peer review. In this, authors deposit articles that are published online without any peer-review. The readers can access those articles freely and can comment on those. In a variant of this model, the journal invites peers to review, once the paper is published online. arXiv, an open access server, publishes articles without peer-review and without any charges. However, it does not allow commenting or review by readers and in essence it is merely a repository.  Even then, this server is quite successful and authors regularly submit high quality articles, particularly in physics and allied subjects. Recently established PeerJ PrePrint archive is an attempt to replicate that for bio-medical sciences. It allows readers to review and comment in articles archived there. Most online journals also allows readers to comment on published peer-reviewed work. However, till now peer review by readers has not catch up the scientific world and very few readers express their opinion online. Even when they comment, those are not detailed like thorough peer-review. Currently, post-publication peer-review by readers doesn't seems to be a viable option. Over-reliance on reader's opinion may also bring  the vices of social media in science publishing. In essence, the current practice of organized peer-review managed by editors is still the gold standard. In the age of open access,  quality peer-review can be sustained only through the author-paying model. Therefore, rather than rejecting author-paying model, we need to develop technologies to reduce the cost of running the show and have to establish some peer-based mechanism to regulate this industry to maintain high standards.

Sunday, October 05, 2014

Why I may ditch the official Android App for Mendeley

In academics, you have to read regularly and have to read a lot. Online access to research articles made our life easy by making full-length research papers a click away. Gone are the days of photocopying and stacking articles in folders.  PDF files of research articles are easy to store, read, and annotate. But ease of access also increases number of papers that one downloads. As the number of papers in the hard disk crosses three digits, it becomes almost impossible to keep track of all those papers and keep those organized. This becomes more problematic if you have multiple interests. A document-cum-reference manager like Mendeley is essential to manage such large libraries. Mendeley is buggy, even after acquisition by Elsevier. Am habituated with its bugs. I use Mendeley to organize papers, with tags, notes and highlights. And am happy. But like many other Mendeley users, I had one long standing demand: to have Mendeley for Android.

In the age of tablets and phablets, users prefer to read papers in their handheld devices. It's very comfortable for me to lean back on a sofa and read a paper in PDF format in my tablet. Highlighting, marking and putting notes, on a PDF file, are easier when you use a touch-screen. Though Android exploded the market for tablets, Mendeley does not have an official Android App. There was lots of expectation that acquisition by Elsevier will lead to the release of one. Though there are news of its development, the App is not out yet. Video and screenshots of a working version of the App were shown in the recently held MDOD14. As per the announced timeline, we may expect the App to be available, for public use, some time in the next year. But even then, I expect that it will have one major limitation.

My network connection is not reliable and it's not always there with me. I prefer to have all the papers of my Mendeley library to be synced and stored in the SD card of my tablet. That's why I have a 32 GB card. This allows me to read these papers anytime, anywhere. Mendeley allows you to store PDF files in their own cloud. It’s free upto 2 GB and beyond that you have to pay. I believe the future Android App for Mendeley will be built around this cloud service. It may even allow one to download files in the device and keep those synced using the same Mendeley cloud. That's where is my trouble.

The size of my Mendeley library, with all PDFs, is already of ~3.5 GB. So I have to use a paid subscription to sync those using Mendeley. I already have a subscription for Dropbox, that I use to sync data files and official documents across multiple devices. Subscribing another cloud does not make sense for me. Additionally, Dropbox gives many options and flexibilities in file management. I want to stick to Dropbox. Therefore, an official Android App for Mendeley may not be suitable for users like me. I want to use the Mendeley but want to sync files using my Dropbox.

Fortunately there is a way out. Referey is an Android App that allows one to view and query a Mendeley database and open associated files stored locally in an Android device. I use the desktop version of Mendeley to create and update my library regularly. Mendeley creates a sqlite database file and keeps associated PDF files organized in a designated folder. I keep this folder and the database file synced with my Dropbox. These files can be now downloaded and auto-synced in my Android tablet using Dropsync. Bingo!! Now I can use Referey to explore my Mendeley database, search papers based on my tags, authors, keywords, journals, year of publication and read the associated PDF files, even in absence of any network connection. As I read those papers, I annotate them and once I have the network, Dropsync syncs those annotated PDF files with my Dropbox.

There are a few more unofficial Android App for Mendeley. But none have the ability to access and read files stored locally. Referey is minimalist and has a very amateurish interface. It does not have any inbuilt PDF reader and you should have one installed separately. Unfortunately, it does not allow editing of the database. That means you can not add any new tag to  a paper. Even then I am happy with it, as it allows me to work offline. And honestly, am not dying to get an official Android App for Mendeley.

Update: Recently purchased a Xiaomi Mi Pad running MIUI, based on Kikat. A tab is perfect for reading papers. Unfortunately, while trying Referey-dropbox-dropsync combination in this tab I met an obstacle. Thanks to Google for making life a bit difficult. From Kikat, Android does not allow third-party Apps to write to anywhere in external SD card. So even though I have a fat extrrnal SD card in my tab, Dropsync can not download all my reference PDF files to my SD card. Also Referey can not read the Mendeley sqlite database file from SD card. Thankfully there are way out. The sqlite file is small. So I kept it in the inbuilt storage of the tab. But I have kept the PDF files synced in the SD card, by a simple trick. The trick is here.