ScienceDirect-Thesisscientist

The Technology Behind ScienceDirect®: Staying Ahead of the Curve

Our previous ScienceDirect® article shared a bit about the history of Elsevier’s platform of peer-reviewed scholarly literature and its evolution through the past 21 years. Today, ScienceDirect offers more than 16 million articles to the world’s research community 24/7/365. This represents some 12,000 GB of data storage. To place that amount of data into perspective, printing all this content on paper would take 600,000 trees! Adding 640,000 articles per year, ScienceDirect data storage increases approximately 200 GB – or another 10,000 trees – each month. Behind the scenes, an Elsevier team is obsessed with making ScienceDirect content available to researchers in a timely manner and a useful format, to better enable research and enhance work efficiency. Using data science and big data, it’s a challenge that they meet and exceed every day.

Concept Preceded Technology in the Initial Development of ScienceDirect

By the early 1990’s, Elsevier recognized that emerging technologies would be pivotal in combining what libraries needed and what it offered as a publisher. As mentioned in the last ScienceDirect article, Elsevier and a select group of U.S.-based universities collaborated on a research project known as TULIP.  Predating the World Wide Web, Elsevier’s technology team digitized two years of its 42 materials science journals, which were then delivered to the participating universities via File Transfer Protocol (FTP). The universities made the journals available to their researchers on campus with their own homemade systems – in essence, each school built its own ScienceDirect.  While the user interface and image quality were far from today’s high standards, researchers were able to view journals online rather than in print format. And with the advent of the web, they didn’t have to be in the library to do so.

TULIP taught Elsevier and the universities two very important lessons. First, publishers needed to create these electronic systems, not libraries. It was not practical for a publisher to provide content to a different system at each university, nor was it cost-effective for universities to create and maintain complex one-of-a-kind systems storing articles they didn’t publish. Second, research libraries needed MORE: content from all disciplines, historical issues of journals going back further than a couple of years, and content from other sources.  It was time for Elsevier to take the lead, becoming the first publisher to develop its own digital library system that could accommodate large amounts of content.

Thanks to a multi-national development team, the first incarnation of ScienceDirect was made available in March, 1997, to 30 research libraries which already subscribed to Elsevier’s journals. With important feedback from these users, Elsevier tweaked ScienceDirect and made it widely available to institutions by the end of that year.

Dedicated Team Maintains and Improves ScienceDirect for New Generations of Researchers

Today, all Elsevier’s journal content, dating back to The Lancet’s first published issue in 1823, is available on ScienceDirect. If those print journals were stacked vertically, they would be taller than Mount Fuji in Japan, or 16.6 Empire State Buildings! With the increased sophistication of technology – as well as those who use it – ScienceDirect also hosts full-text content from tens of thousands of books.

Who’s in charge of ScienceDirect and this massive amount of content? SD Tech, a part of Elsevier’s Technology organization. Located mostly in London, UK, SD Tech designs, develops and tests the features and functions that make up ScienceDirect, from personalized search alerts to topic pages. The platform team builds the software, tools and devices necessary to support the testing and releases, and also maintains ScienceDirect. SD Tech is nearly 80 people strong, and includes software engineers, quality assurance (QA) testers, data and business analysts, user experience (UX) specialists, product managers, and a management team.

Users Are the Top Priority in the Evolution of ScienceDirect 

SD Tech’s User Experience team tests concepts, prototypes, and live products with real researchers during all phases of the design and development process. They look at usefulness, ease of use and delight. In 2017 alone, the UX team conducted 23 user experience studies. They interviewed hundreds of researchers, both in person and remotely, and surveyed an additional 550 researchers, librarians, and other users online.

Diligent Software Development and Testing Keep the Technology Updated on All Platforms

Using state-of-the-art software development methodologies, the ScienceDirect development teams can introduce new features or functions multiple times in one day. This eliminates development delays, allowing researchers to benefit from these improvements very quickly.

All new code goes through rigorous testing before it’s released on ScienceDirect. This results in a smooth and user-friendly experience, as well as the elimination of potential bugs. Knowing that researchers access ScienceDirect on multiple platforms in various locations, testing is performed in Chrome, Firefox, and Internet Explorer, using both macOS and Windows. Elsevier also performs tablet and mobile technology testing in Safari and Chrome on iPads, iPhones, and Androids. This promotes seamless access to the ScienceDirect content by all potential users.

In 2017, SD Tech successfully completed more than three thousand individual software releases to ScienceDirect – at least eight per day. That rate has doubled this year, with nearly 2,000 releases in only the first four months of 2018. With an ever-increasing number of ScienceDirect users, Elsevier tirelessly maintains the stability and functionality of the platform.

Performance is Closely Monitored for Optimal Service and Discoverability

Elsevier realizes that due to the lightning-fast pace of scientific discovery, it is more critical than ever for researchers to find the content they need without delays. Monitoring and alerts play a major role in sustaining the performance of ScienceDirect. SD Tech maintains two distinct geographic regions, to help balance the usage load on ScienceDirect. If one region experiences reduced performance, technology automatically reroutes traffic to the healthy region. This eliminates slow response times for users during peak usage hours. The platform team runs continuous load testing in QA, and Web services are monitored 24/7. If an issue does arise on ScienceDirect, SD Tech is alerted, and the issue is isolated and resolved quickly.

Underlying Technology Leads to Outstanding Results

Thanks to the dedication of SD Tech and the sophisticated technology they employ, most sources rate ScienceDirect as one of the most popular sites on the Internet. It is a time-saving tool like no other: ScienceDirect users retrieve documents at a rate of 5.2 seconds per article page and 2.8 seconds per search result page. This is particularly impressive considering that many documents on ScienceDirect include embedded multimedia and supplementary content.

Machine learning – a subset of artificial intelligence – provides ScienceDirect users with the only personalized recommender based on a major full-text research content platform. The personalized recommender finds the right information for each researcher by automatically recommending relevant content without requiring the user to explicitly enter a search query. Algorithms first determine researchers’ interests based on their usage of ScienceDirect. Then by applying machine learning-generated models to those usage logs, ScienceDirect captures not only the obvious relationships but also reveals more subtle ones…such as discovering that two articles from different domains that use different terminology are actually related to one another. By applying user data and Elsevier technology to the problem of information overload, researchers can find more relevant research and spend less time looking for it.

Elsevier has an unwavering commitment to investing in technologies that can upgrade the ScienceDirect experience. Just this year, by moving to a new search and retrieval service, and becoming fully hosted on Amazon Web Services, ScienceDirect users will see improvements in flexibility, performance and availability.

Next Up:  Brand New Journal Homepages on ScienceDirect

The next ScienceDirect article will introduce the next generation of journal homepages. Replacing the existing journal pages on ScienceDirect, the revamped homepages increase usability by combining richer information, clearer design and faster download times. With ScienceDirect users in mind, the pages have been created to integrate with other Elsevier platforms where journal content appears, contributing to a more seamless, personal and social experience.

By: , Posted on: July 27, 2018

Elsevier