April 2012 edition!

     The Ranking of Repositories is published since 2008 and two editions are available usually at the end of January and July.

THIS IS AN EXTRAORDINARY EDITION INCORPORATING A FEW MAJOR CHANGES IN THE METHODOLOGY, ENLARGING THE DATABASE WITH 300+ ADDITIONAL REPOSITORIES AND CORRECTING SEVERAL MISTAKES

     The repositories should have their own web domain or subdomain and include at least peer-reviewed papers to be considered (services that contain only archives, databanks or learning objects are not ranked).

     Although there is a paper describing the Ranking,

Aguillo, I.F., Ortega, J.L., Fernández, M., Utrilla, A.M. (2010). Indicators for a webometric Ranking of Open Access Repositories. Scientometrics, 82 (3): 477-486. Click here to get the author's edition in Open Access.

there are some important changes in the methodology that should be taken into account:

     Published figures are RANKS (lower is better), intended for showing individual performances, but they are not the values used in the calculations.

     Link analysis is a major component of our composite index. Unfortunately not the successful PageRank, the Google algorithm, nor the numbers of backlinks provided by the now defunct Yahoo Site Explorer are publicly available, so a new data provider was required.

     The visibility is obtained from combining external inlinks and referred domains according to Majestic SEO (50%).

     Due to the different problems with the stability of most the public search engines, it was decided finally that only Google will be used as the source for all these indicators (new weights starting April 2012 edition):

• Size: Total number of web pages (by Google), excluding the rich files (10%)
• Rich files: The sum of the following file types(by Google): pdf, doc+docx, ppt+ppts and ps+eps (10%)
  
• Scholar: The combination of the total number of papers in Google Scholar and the recent values for the 5-year period 2007-2011 (30%).

Registration of repositories with Webometrics

     Please, remember we are not asssociated with directories like OpenDOAR or ROAR and that our criteria are more restrictive. If you wish to add you repository to our Ranking you should contact directly with us informing of the name and URL of your repository in the following e-mail address.

Suggestions from the Ranking Editors

     The role of librarians in the Open Access movement has been very important as they are the main designers, promoters and maintainers of institutional repositories. The basic principles behind the current arrangement of repositories are inspired by the librarianship techniques.

     However, some decisions guided by these principles are probably hampering the universal adoption of the Open Access initiatives (OAI), as they are ignoring the main users of these repositories: scientists, researchers, scholars.

     We publicly acknowledge the work done by librarians worldwide but we would like to offer some suggestions to improve the impact of the OAI:

• The institutional repository is a very important asset of the institution as a whole, not only of the library. We recommend the following syntax for the institutional repository web address:

  http://repository.university.country/
  and not
  http://repository.library.university.country/

• Collective catalogues are well suited for library purposes, but although harvesters and consortium repositories may be very useful, they should be considered always as secondary systems, mainly for purposes of increasing visibility and searchability. Even for very small institutions it is feasible to set up their own repository that is, as mentioned before, the main research asset for the institution.
• Repositories consist of full text papers, not bibliographic records. That means that the link to the full text file should be clearly marked in the web page of each record. Some repository designs makes difficult to locate such links.
• For scientists it is important that the link to the full text would be easily citable. So it should be avoided the following situations:
  • Very long URLs.
  • URLs with nonsense characters that have no meaning. We suggest to use, for example, the last name of the first author and the publication year.
  • File names without suffix indicating the type (.pdf, .doc, .ps, etc). You can follow the recommendations indicated here and here.

  As usual we will thank your comments if you address them to our MAILBOX .