[BOAI] Re: Gratis vs. Libre OA

[Andras Holl]

Hi,

This is a delayed comment on a posting by Stevan Harnad, dated 18th december, 2011. I was glad to read that gratis OA is not second class OA after all.

Andras Holl

Is libre OA and CC appropriate forscientific literature?

The goal of this essay is not toquestion the usefulness of OA - OA is not only good for science and society, butnecessary. I argue in favor of the gratis OA model, or a readjustment of theconditions of OA, for creating a copyright model more suitable for scientificarticles than CC, and for the creation of a protocol which is capable oftransmitting harvesting semantics, access and re-use rights in a wayunderstandable of machine agents.

Gratis vs. libre OA

I find libre OA unnecessary. While OA(gratis OA) is indispensable, I do not see real demand for libre OA. I cancook up cases when libre OA would be advantageous - and I can give a fewexamples when it proved to be harmful. But the main point is that libre OA isnot necessary - while OA is.

Who has the rights to reposit?

To achieve OA, the green road is openfor everyone. While I do not see much reason to deposit articlesappeared in OA journals to repositories, I do not object to that. Non-OA articlesshould be made OA with depositing them to repositories. But who can do this?Obviously those who have rights over the article. The author(s), theiremployers, and the organizations involved in funding the research. In my view, noother parties are entitled to re-distribute the article. In other words, the wholejournal, or whole volumes would not be redistributable normally, but eacharticle individually would (by the parties listed above).

Copyright models

There are several copyright formulas,each designed for a certain kind of documents. GNU or BSD is for software,CC is for artwork. I do not think that these are appropriate forscientific journal articles. I can understand why is there a demand for re-using softwarecode or executable programs, or a nice illustration. As a scientist, Ithink that "fair use" and OA conditions proposed above (gratis OA, noredistribution for others than who could claim rights over the article) are enough forscience. Or maybe we should specify something better tailored toscientific journal articles.

Article components

While I do not see any demand for wholearticles to be freely redistributable and reusable or remixable - only for getting them accessible freely,  I do see much demand forsome building blocks of the same articles. There is much demand for re-usingfigures. Scientists often want to reproduce figures from published articles,unchanged or modified, plotting new measurements, removing questionable data points.There is also a great demand for re-using research data. One scientist might want to plotthe data from the literature together with his/her own, or there are compilation databaseswhich collect data from the literature. The demand for mashups might not be large as yet,but I think it will grow. In these cases redistribution and re-use rights are of greatimportance. Here I think CC by-nc could be used. Another redistributable component - somethingeven desirable to get redistributed - is the meta-data, includingthe abstract.

Resources for communicating thesemantics, hints for harvesters and machine-readable rights

While it is necessary for scientists to get access promptly to an article they need - something what is provided by gratis OA, mass harvesting of journal content is not usually done manually. I can not imagine a situation where it would become suddenly important, from one moment to an other to mirror a whole journal. I feel that such situations could be handled without providing libre OA rights to everyone. On the other hand, some parts of the articles - the abstract, the figures, the data - should be made available with well defined conditions, in advance. But the demand for these article components might come not only for individual scientists reading the article, but by robotic agents.

I feel that we would need a mechanism -maybe a modification or combination of existing ones - to enable crawlers, harvesters to fetchwhat is the best for them and for the journal or repository, and to communicate the rights forautomatic creation of mashups. Existing technologies like OAI-ORE, robots.txt and .htaccessalready have much (or maybe all) what needed. I illustrate my idea with an example. OAI-ORE coulddescribe the semantic relationships between the files on the journal website (OAjournal) or in the repository. This is an article and that is thefirst figure of it, and that is the first table. All items could bepresent in multiple formats: the same article could be available inHTML and PDF. While it is obvious that PDF is for printing, HTML isfor on-line viewing, the additional information that the HTML offersenhanced features could be communicated too. One might want to give the hint to aharvester or indexer - a machine - that for text indexing the recommended format
  is the LaTeXsource. This information might be present article-by-article, but itwould be even better to inform the harvesters or indexers about thestructure of the website: where to find material to index, and whichfile is equivalent text-wise to what, based on file-names ordirectories. The copyright information should be also communicatedwith the robots (with the human reader too, maybe by other means).The crawler should be able to understand that which information couldbe re-used in mashups, for instance.

Andras Holl

-------------------------------------------------------------------------------- 
Andras Holl / Holl Andras                e-mail: holl at konkoly.hu 
Konkoly Observatory / MTA CsKI           Tel.: +36 1 3919368 Fax: +36 1 2754668 
IT manager / Szamitastechn. rendszervez. Mail: H1525 POBox 67, Budapest, Hungary 
--------------------------------------------------------------------------------

 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ecs.soton.ac.uk/pipermail/boai-forum/attachments/20120121/a82cb881/attachment.html