Re: DOI

From: Ed Pentz <epentz_at_CROSSREF.ORG>
Date: Fri, 16 Feb 2001 17:24:48 -0500

I wanted to reply to some of the issues about CrossRef and DOIs raised in
this interesting discussion. CrossRef's purpose is to make scholarly
journal article reference linking efficient, scalable and persistent. In
order to do this there needs to be a linking infrastructure of identifiers,
standardized metadata and a resolution system to get to the content.
CrossRef and DOIs provide this infrastructure.

Leslie Chan wrote that:
>>[CrossRef's] interlinking schema is an exclusionary one in that unless
>>you are a member of the CrossRef group, your publications will not be
>>interoperable with theirs. As CrossRef is dominated by large
>>multinational publishers who hold a near monopoly on science
>>publishing..."

CrossRef now has 68 member publishers and 42 (~62%) are non-profit, some of
them very small. In order to link references publishers have tended to
sign separate agreements with each publisher they link to and have used
algorithmic URLs (ISSN, volume and page) for links. This is labor
intensive and doesn't scale very well - it's not easy tracking the linking
schemes of potentially hundreds of publishers. With CrossRef and DOIs,
publishers can link to, and be linked to from, all the other CrossRef
members. CrossRef is a quid pro quo and value goes both ways. A member
must deposit metadata for its articles in the CrossRef system (CrossRef has
no publications or full text itself, so can't determine if publications
will be "interoperable") and a member must link to all the other members.
Participating in CrossRef does not stop publishers continuing to use their
algorithmic linking schemes or using something like OAi. Publishers have
been joining CrossRef because it makes it easier for them to link and for
others to link to them.

Stevan Harnard wrote:
>>CrossRef is a proprietary reference linking service between journal
>>publishers. It is open to data providers/publishers who want to pay. In
>>this sense it is perhaps less discriminatory than DOI-F. It is an
>>implementation of the DOI.

CrossRef is a collaborative linking service for scholarly publishers.
CrossRef charges fees and is an implementation of the DOI. Publishers pay
to deposit their metadata and register DOIs ($.60 per article for 2001
content, $.10 for 2000 and before) and there is a $.10 fee for each DOI
retrieved. Retrieval of DOIs is open to Affiliates (secondary publishers,
abstracting and indexing databases and libraries, who pay a flat $500 per
year for unlimited retrievals.) The DOI Retrieval Fee is a one time fee
for looking up the DOI - once someone has a DOI, there are no charges at
all for creating or clicking DOI links. Think of ISSNs, ISBNs and bar
codes - these all charge fees and have managed systems to control the
assignment of identifiers. CrossRef fees are in relation to use
(depositing and retrieving DOIs, not use of DOIs in links), so a small
publisher with a few journals pays a small amount.

CrossRef does not collect abstracts or full text and it is 'business model
neutral' - publishers determine and control access to their full text
articles. The full text can be free or it can be limited to subscribers
through IP authentication or other means.

The linking infrastructure that CrossRef and DOI provide is very important
to the future of linking. If we want more sophisticated linking, such as
localized linking (i.e. appropriate copy linking solutions like SFX),
multiple resolution (one DOI going to multiple URLs) and other services,
managed systems of identifiers and metadata are necessary. CrossRef and
DOIs won't be the only options, but for scholarly journal publishers, they
are the main option.

CrossRef is working with the Digital Library Federation, DOI Foundation,
Los Alamos, Univ of Illinois, Ohio State University, SFX/Ex Libris and
others on a practical prototype of localized linking incorporating DOIs,
OpenURLs, SFX and other local linking servers - (see a report of the July
workshop at http://www.niso.org/CNRI-mtg.html).

Regards,

Ed Pentz
Executive Director, CrossRef
epentz_at_crossref.org


On Fri, 16 Feb 2001 09:38:19 +0000, Steve Hitchcock <sh94r_at_ecs.soton.ac.uk>
wrote:

>At 09:16 12/02/01 +0000, Stevan Harnad wrote:
>>
>>Your interpretation is correct. DOI is proprietary, OAI is open.
>>See: http://www.ecs.soton.ac.uk/~harnad/Tp/resolution.htm
>
>In the context of the debate below there are a number of subtle
>distinctions that might be worth making to avoid confusion. Also, it is
>important to define what 'open' applies to.
>
>In one sense, DOI is no more proprietary than OAI. It is based on an open
>technology, the Handle server used to resolve the DOI identifiers. Anyone
>can build an ID scheme based on the Handle server
>
>DOI is a particular ID scheme administered by a 'not-for-profit membership
>organization', the DOI Foundation. True, it represents the interests mostly
>of large commercial publishers, but the principal barrier to participation
>is the DOI-F's wish to preserve the integrity of the scheme and DOI-based
>resolver services by restricting those who can assign DOIs to works and who
>can operate resolvers.
>
>The OAi makes metadata available to describe works, but doesn't mandate
>access to works. In this sense OAi is more open to service providers, those
>who collect data from data, but not from the point of view of all data
>providers and users.
>
>CrossRef is a proprietary reference linking service between journal
>publishers. It is open to data providers/publishers who want to pay. In
>this sense it is perhaps less discriminatory than DOI-F. It is an
>implementation of the DOI.
>
>Neither DOI nor CrossRef 'block' access to works. They improve access for
>those who are able to pay. The 'blocking', i.e. the price that has to be
>paid to access journals, is already in place. There is an idea that DOI
>will track usage of rights-based materials to prevent unauthorised copying.
>Maybe that's what is meant by the claim below that "the underlying goal of
>the initiative is essentially to block access to the referred journal
>literature".
>
>OAi doesn't have an 'associated open citation linking mechanism', yet. One
>distinction that needs to be made about 'open' linking is whether open
>applies to the software that applies the links, to the ability of the
>format to present the links, to the ability of the user's service to
>provide the links, or to the ability of the user's service to retrieve a
>document using a link. I'll assume that this correspondence is concerned
>with the latter two.
>
>In this respect it is true to say that CrossRef linking is 'exclusionary',
>but it doesn't necessarily follow that unless you are a member of CrossRef
>'your publications will not be interoperable with theirs'. It is possible
>that publishers will make information about journals and papers available
>in OAi tagged form, perhaps using Eprints.org software because it is
>convenient and cost-effective. This provides a degree of interoperability,
>and doesn't directly provide links or access to works, but could be the
>basis of both. It is certainly in publishers' interests to be open and
>interoperable at this level at least, whether participants in CrossRef or
not.
>
>An additional couple of points on reference linking services and systems.
>
>SFX is a proprietary system and not the basis of OAi (it's concerned with
>the 'appropriate copy' problem, which is a library issue rather than the
>global OAi issue). Herbert Van de Sompel, one of the developers of SFX, is
>however a primary mover in OAi, and so a better paper to start with is
>The UPS Prototype: An Experimental End-User Service across E-Print Archives
>http://www.dlib.org/dlib/february00/vandesompel-ups/02vandesompel-ups.html
>
>The original jake is a database of journals rather than a reference linking
>service, although a version of it is now available through an Openly Inc.
>server as a linking service
>http://jake.openly.com/
>Openly's strategy is to design truly open environments in the hope of
>stimulating use, and then offering the best tools and services for the
>environment they created.
>
>This is an increasingly complex area, but I think it is important to
>identify the objectives of an initiative to understand its context, instead
>of applying labels that might be inadvertently pejorative and obscure
>rather than clarify.
>
>Steve
Received on Wed Jan 03 2001 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:03 GMT