arxiv.org (pronounced like the word ‘archive’) is a beloved 30 year old website. Yes, you read that right! 30 years ago there wasnt really even a web yet; arXiv was birthed during a very early iteration of the internet while discovering new uses for something called “email”.
From those early days, it was rapidly adopted by scientists across Physics, then Math, and, over the last decade Computer Science. It is not just a website to it’s users, many researchers have their entire life’s work hosted on the site. They have contributed to it’s sucess through use, feedback, word of mouth, and as volunteer moderators. It has always had a co-created spirit, scrappy, rather clunky, but familiar and trustworthy. In that environment, change is not necessarily welcome when it rears it’s head.
But the need for change has come. The old website is out of compliance with accessibility mandates, not mobile friendly, and is inefficient to both use and to keep updated. What is needed is a modern codebase, embracing accessibility and usability, easy to maintain, and that can flexibly adapt to changing content needs. In short, arxiv.org needs a redesign.
The redesign spec
The redesign spec includes a number of important stakeholders, with sometimes conflicting needs. For example, our funders want more visible representation while the majority of users would read that as selling out; Veteran users do not see the need for change, while younger users and new stakeholders clamor for more functionality; Lastly, Cornell University, our stewards, require the website redesign to reflect their brand, but my UX research shows that scientists around the globe would see an emphasis on just one University as a sign of dangerous bias in publishing.
Our redesign threads this ever narrowing needle eye with two strategies: A spare visual design that foregrounds the functionality our community has been asking for, and a change management plan that slowly unfolds to the public. The generous timeline allows for psychological adjustment and multiple rounds of feedback on a beta site, and a great deal of communication from my team as well as outside voices from different scientific communities, all before a single new change goes live.
To further limit the amount of disruption, planned content changes will be part of a future phase. In this phase we will focus on an accessible framework and improved efficiency. Though it slows the timeline significantly, we are moving forward with respect for all of arXiv’s stakeholders, old and new, and striving for measured balance in our rollout strategy.
In with the new…
In the examples below the new layout is on the left, contrasted with the legacy website that users know so well on the right.
The new header not only increases vertical density, it adds a navigation option for the first time in arXiv’s 30 year history. The navigation, like the overall design, is spare and reflects the priorities of our users. Minimal branding is included for Cornell at the top, while all funders branding has been put in the footer. Login and account links have been standardized. Cornell brand tie-ins were achieved through font use and minimal ‘Cornell Red’ highlighting, keeping the brand focus squarely on arXiv and avoiding the dangerous perception of bias.
The search page clarifies content and simplifies the interface…
..while search results are more legible and increase vertical density.
The mobile layout… exists now!
By carefully listening to our community during all stages of design and respecting their long history and rich investment in the website as it is now, we have created a redesign that threads the needle. Though the change management process will not complete until Spring 2023 it is on a strong footing for success and community acceptance.