Should the journal article change, and if so, how? In this multimedia age, the workforce is increasingly populated by people who grew up with the internet, scholarly publishers anticipate the demise of the traditional article and spend lots of time rethinking how best to present the information it contains.

Elsevier, for example, first developed its “article of the future” prototypes for Cell Press journals in 2009, and last year unveiled three further ones, available in seven disciplines.

The “big bang” approach isn’t necessarily needed. Articles evolve over time. Some publishers, for instance, have dispensed with “web extra” data supplements. Many now routinely embed videos and audio files, include article level metrics, provide peer review comments as part of its pre-publication history. Others serve mobile-optimised versions of articles because of the growing number of accesses from smartphones and tablet computers, and link to author blogs. Social media and bookmarking and research management links are now standard features on many articles.

HighWire Press, web hosting partner of the BMJ and more than 1600 other academic journals, regularly surveys readers, authors, and editors, to see how their reading habits change over time, and how they like to receive information.

Its latest findings are based on interviews with 45 Stanford University researchers (excluding students and teaching staff) and a separate discussion on the best ways to communicate science with 16 biomedical researchers and clinicians.

HighWire also hosted separate discussions with 19 postgraduate students, journal editors, publishers, librarians, and technologists. The theme of the meeting was: “What should or will change in research communication?”

In a nutshell, according to the feedback provided to HighWire, blogs are for entertainment, not to advance science. HTML is ugly to read (and used to browse, unlike pdfs, which are printed off mostly and saved for reading later). Journal articles are a static way of publishing that the next generation won’t be happy with, and tend to get read in their entirety only if a reader sees his or name cited.

One respondent said he didn’t read journals, he searched databases. Another asked for audio abstracts because it’s quicker to listen than to read.

HighWire regularly seeks the views of the research community, and one consistent theme seems to be that for most, mobile access at work is via a laptop rather than a mobile phone.

In many ways this is understandable. Full text research articles can be long (the BMJ’s, for example, have no word count limit, although a shorter “pico” version appears in print). All of the features described above can clutter an article, making it harder to read on a phone.

But traffic to bmj.com from mobile devices (including tablet computers) is rising fast. Perhaps this is because the journal’s core readers are practicing clinicians, and if it is to deliver on its mission statement (helping doctors make better decisions), many access the site in the course of their working day while they are on the go. Also, the journal publishes not only peer reviewed original research and education, but shorter comment and news items that are easier to read on mobile devices.

Interestingly, we tend to present all articles (scholarly and journalistic) the same. Perhaps the question for us is whether we should.

David Payne is editor, bmj.com.