A Step-by-Step Approach to Content Management

lemurLoginIllustration

Image credit: Michael Morris

Today at PLOS, we celebrate an important milestone: We launched the first iteration of the new PLOS CMS (codename: Lemur). This first installment is a homepage editor for six of the PLOS journals (Biology, Medicine, Pathogens, Computational Biology, Genetics, and NTDs). Our homepage editor is a browser application that facilitates curating and preparing content to feature on the journal homepages. It lets our journal staff queue up items to feature, write blurbs about each item in a WYSIWYG editor, select images, and perform basic image manipulations. It has a drag-and-drop interface to easily reorder featured content, and as you’d expect, it includes previewing and publishing controls.

Our launch timing, not so coincidentally, corresponds to the launch of our updated homepage design for the aforementioned six journals. The new, more sophisticated homepage design called for sophisticated curation controls, so it was an easy call to make the homepage editor the first functional area for the new PLOS CMS to tackle. The old process for homepage updates was manual and restrictive. But we’re restricted no longer—journal staff now have all the controls they need at their fingertips, without having to be HTML experts.

PLOS’s new CMS is an internal tool… for today. But that could evolve over time, as the scientific publishing community moves to a brave new world that’s “beyond the journal”. I’ll dig deeper into the idea of meeting an evolving landscape by answering the question you’re probably asking yourself right now:

Why are we building our own CMS?

There are several categories of answers to this question, presented here in no particular order:

  • Flexibility: We could have gone with an off-the-shelf CMS. But it’s likely that we would have actually required two different off-the-shelf content management systems: one to manage our web content, and one to manage our article corpus. These two content types generally utilize two very different kinds of CMS. The idea of adopting—and adapting—two different CMSes (or select just one type of CMS and heavily customize it to work with both content types) was not very appealing. This is a particularly unappealing solution when viewed through the lens of our ultimate goal, which is to blur the lines between all of our different content and media types, and interact with and present any combination of them seamlessly, in ways we can’t necessarily predict today. That’s why we have opted to separate the functions of content management, and build a curation layer for pulling the content together in the ways we need.
  • Innovation: We don’t want to miss opportunities to innovate on content delivery. We suspect that innovation would be slowed if we become locked into monolithic systems that are difficult to customize.
  • Technical considerations: Separation of content management functions from a technical perspective makes a lot of sense. You hear of a variety of publishing operations opting to roll their own for the same reasons, including the New York Times. This separation should also make us more agile in the sense of being able to respond relatively quickly to emerging needs, new ideas for innovation, or other interesting technological developments we want to try out.
  • Longer term community ideas: We start today by building Lemur with specific curation tasks in mind. But it’s easy to imagine a future in which anyone could curate PLOS (and other) content, using tools we provide. Imagine societies or educators using our content and curating it to their own needs, which may or may not be tied to the traditional model of the journal. Many examples of how this could evolve spring to mind!

Lean makes it better

We’ve been using Lean methodology to develop the content management system, component by component. We’ve been testing our wireframes throughout the design process. We’ve co-located the entire team in a conference room—since February, mind you! We’re reaping the benefits (and enjoying the process) of pair programming. Close collaboration among developers, UX, and product owner noticeably improves both our product and our velocity, while collective code ownership ensures maintainability. And here’s the part that requires lots of discipline: We release a minimum viable product, so we can gather observations of how it is used and see how it could be improved, rather than anticipate the entire product in a vacuum. As we improve the individual features with the findings from our observations, we will also incorporate that learning in the approaches we take with the rest of the system.

 

This entry was posted in Tech and tagged . Bookmark the permalink.
Add Comment Register



Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>