Products & projects
Can't display this module in this section.
Can't display this module in this section.

What should a modern research publishing platform look like?

After recently complaining about the state of academic publishing, I started thinking about what could replace it. Here are my thoughts on features such a service would have.

In my previous blog post I identified 6 criteria by which we could judge a good research publishing platform:

  1. There should be an efficient way to discover material that is relevant to our interests;
  2. There should be a filter for quality, so that we can spend our limited time reading worthwhile papers;
  3. There should be a way for “outsiders” to know whether the research is generally accepted by the research community;
  4. Research should be available for public consumption as soon as possible after submission;
  5. Anyone should be able to access the research, at no cost;
  6. Anyone should be able to publish research, at no cost.

When designing a system to meet these criteria we should add two extra conditions, in order for it to be successful:

  1. It should provide value to users right from the start, even if no one else is using it yet;
  2. It should have a “network effect”, i.e., it should get more useful the more people who use it.

Incidentally, too many internet startups fail because they don’t meet both of these criteria. An idea that would be great if everyone was using it, but only useful if everyone is using it, will never get traction.

The system I’m thinking of is a web-based site that anyone can access. Researchers can create a project for their research, and interested users can “follow” updates to the project (like on github). It can be a place where researchers can collaborate, planning and brainstorming ideas, either in private or in public. It could be a central place where debates on the research could occur. Any “papers” that are produced can be hosted there, like a pre-print service, with permanent links to enable proper citation. Any data or code associated with the research could be hosted there, so that everything is in one place, with source and version control. A researcher could have a private (or public) “field journal” section, like a blog, where results, graphs, notes, links, etc. could be placed for easy searching and access. I think that this sort of resource would be useful for researchers right away, and could initially be used in parallel with regular publishing.

Users should be able to “like”, or “vouch for” a particular researcher, perhaps only in the context of a particular field, and the same for a particular project. Perhaps algorithms could be developed to create a score for each project, based on who has “liked” it and what their reputation score is. It would need to be more than simple voting, so that you don’t get the ignorant masses voting up their favorite topic. This type of reputation and scoring is difficult, but not impossible. Klout seems to have a reasonable algorithm for social network influence, and Hacker News seems to have some special way of calculating their scores for surfacing content.

The last few years have made apparent that “the crowd” on the internet is good at surfacing interesting content quickly; this should apply to academic research also. A good search system, with appropriate tags, and flexible alerts and saved searches should make it easy to find relevant information. I’m imagining a service that notifies me via my chosen channel (RSS, email, iPhone notification, etc) for any of the following conditions:

  • Anyone publishes a paper with “tabu search” in the title or keywords;
  • Any of the researchers on my “white list” publishes anything;
  • A project I’m interested in posts an update of any kind;
  • Someone posts a comment on one of my research projects;
  • A paper has been published in the “operations research” field, and it has achieved a score of at least 35 (made up number for this example);
  • One of the researchers I trust as a content surfacer has shared a link to an interesting paper.

A flexible enough system of search options (I’m even thinking regular expressions!) could be very powerful.

We could get even more ambitious and try to tackle funding for research as well. Kickstarter is a site where people can post a project that they want to do, perhaps create a product or shoot a movie, and ask for people to help fund it. Users commit a certain amount of money to the project, and this is only actually claimed if a predefined funding threshold is met, so you know you aren’t wasting money on a project that will never get off the ground. This site I imagine could be a Kickstarter for research. I’m sure that lots of people would be willing to contribute money to research they believe in. Crowd-funding looks to me like the future.

Anyway, that was a bit of an idea-dump. Is this already being done? What am I not considering? If it sounds like a good idea, then who wants to do it?

PrintView Printer Friendly Version

« I am a statistician and I buy lottery tickets | Main | Academic publishing 2.0 »