Archive for the ‘normalization’ tag
It is now two years since the ticTOCs Best Practice Recommendation group, headed by CrossRef and consisting of members from Talis, Nature Publishing Group, Oxford University Press and Heriot-Watt University; published the “Recommendations on RSS Feeds for Scholarly Publishers.”
RSS feeds are designed to be aggregated and reused by other services and software applications. In general RSS feeds should always be created with this in mind. The Recommendations are in full agreement with this principle.
Back in 2009, two practices were noticed by the ticTOCs Project:
- there was a wide variation amongst the journal TOC RSS feeds produced by scholarly publishers, and
- in most of the cases the feeds’ content had very limited information on the articles, such as uniquely the title and the link to the article’s webpage.
Variations in the way publisher implement RSS feeds basically preclude the consistent and automated aggregation of feeds. At the same time, having little content to offer, limit the reusability and value of feeds for other services that want to create interesting applications by combining the feeds. The Recommendations were created to help publishers avoid the inconveniences created by those two practices, and to advocate good practice in the production and provision of TOC RSS feeds for scholarly journals.
There are signs that the Recommendations are gradually being embraced to a certain extent, but how many scholarly publishers have really implemented the Recommendations in their journal TOC RSS feeds? There’s no way to get an exact number, but we can get a good idea of the progress being made by taking a look at the number of journals that are using the four RSS 1.0 modules recommended by the group, namely Admin, Content, Dublin Core and PRISM modules.
Today we have examined the RSS feeds of the journals collected by JournalTOCs to get an approximate picture of how many publishers are making the move. Currently 17,112 journals from 917 publishers are being indexed by JournalTOCs.
Interestingly no journal uses the Admin module in their RSS feeds. Only a few hundreds of subscription journals make use of the Content module. However those two modules are not particularly relevant from the re-usability perspective (the Admin module is intended to be used by consumers of a feed to provide feedback on errors encountered in the feed and the Content module is used to include formatted HTML marked up content for browsers.) The modules that really can give us a good indication of the Recommendations’ uptake are the Dublin Core and PRISM modules.
8,025 journals are using Dublin Core, PRISM or both modules; but only 3,673 of those journals are using both modules.
If we put the figures from the number of publishers’ perspective, 425 publishers are using Dublin Core, PRISM or both modules; and 295 of them use both Dublin Core and PRISM modules.
Regarding Open Access Journals, there are 2,660 Open Access journals in JournalTOCs, and 708 of them have implemented either the Dublin Core or the PRISM module; but only 288 of Open Access journals use both Dublin Core and PRISM modules.
In conclusion: There is still a long way to go. Only 31% of the publishers are using the two main modules and in some extend have adopted the Recommendations. This is equivalent to 22% of the journals. To make a real progress two things should happen: (1) Elsevier, Springer-Verlag and Taylor and Francis together publish over 6,000 journals. A significant step forward will only be made when those three large publishers adopt the Recommendations. (2) An inexplicable low number of Open Access journals have implemented the recommendations. Without proper orientation and guidance, the publishers of OA journals so far haven’t been able to grasp the benefits of adopting best practices and using standard modules for their RSS feeds.
The affiliation of an author is vital information for a successful service such as the journalTOCs API. So far, the most common query received from our users involves this request: “find the papers published by authors from my institution”
Taking into account the Recommendation # 6 of the “Guidelines for Encoding Bibliographic Citation Information in Dublin Core Metadata” proposed by Ann Apps, MIMAS, the journalTOCs schema has been upgrade to include the
dc:contributor element to capture authors’ affiliations.
Thus, the normalized bibliographic metadata schema that JournalTOCs is using for syndicating journal TOC metadata in RSS format includes now the elements shown in the following figure.
Basically, the reason behind using the “contributor” Dublin-Core element is that the affiliation for a journal article pertains to the resource. In this way we also recognise the contribution made by the author’s institution to the creation of the resource. Because affiliation is a property of the resource, rather than a creator, the fact that there is no way to explicitly correlate particular authors with their affiliations shouldn’t be an issue. Similarly the fact that the affiliation will no longer corresponds to the author when he moves to a different institution shouldn’t negatively impact to the current awareness service provided by the journalTOCs API.