CoNE Journal

Preliminary descriptive Metadata

current metadata configuration For the selection of the descriptive metadata the main focus has been set on the minimum level of information that is needed to disambiguate entities. The list of descriptive metadata elements is extendable by new elements.

Metadata elements:

• Journal title [1]

The name of the journal (e.g. "Journal of the ACM")

• Alternative title [0-n]

Any alternative name or abbreviation of the journal

Remark Inga: Tagging of abbreviations as such? Indicating the origin of abbreviation if known? Discussion: see below

• Publisher [0-n?]

The name of the institution that publishes the journal

• Identifier [0-n]

Any external identifier (e.g. ISSN, EZB-ID, ZDB-ID)

Schema has to be indicated

• Locator [0-1?]

Locator of the authority file source

Question Inga: Do we mean an URL pointing to the record?

• Rights [0-n]

Statement on open access availability

Discussion: see below

• Subject [0-n]

Subject/domain field of the journal

Possible Relations:

• isSuccessorOf
• isPredecessorOf

Rights statement for journals

Update on <rights>: as there is no requirement from Christoph/Anja for rights statements on journal level, we can choose whatever provider/Whatever information. I would opt for DOAJ, as it gives at least clear indication, which journals are OA, although no information on "Green" road publishers. disadvantage romeo/sherpa: indicates on publisher level, but not on journal level.--Ulla 16:16, 11 January 2008 (CET)

Requirement: The information collected under PubMan OA Statistics provide no clear picture for what kind of request the right statements are required. Is the goal either to receive the information if specific articles are open access or if the journal supports oa-publishing (for all articles? for some articles? via author-pays model?). We probably should avoid to include rights information until we have a clearer picture. --Inga 17:38, 18 January 2008 (CET)

Values: How do we populate i.e. what value has the rights metadata in the journal if the journal is OA (in accordance with DOAJ? (there are statements like: http://www.doaj.org/doaj?func=loadTempl&templ=faq#definition)--Natasa 15:41, 18 January 2008 (CET)

An overview on OA levels is provided in the wikipedia article on Open access journals:

Level	Explanation	Value	Source
1	Journals entirely open access	gold	DOAJ
2	Journals with research articles open access	??	no source
3	Journals with some research articles open access	??	no source
4	Journals with some articles open access and the other delayed access	??	no source
5	Journals with delayed open access	??	no source
6	Journals permitting self-archiving of articles.	green	sherpa/RoMEO

Some thoughts on DOAJ

1. DOAJ is a directory of open access scientific and scholarly journals. Each month new journals are added and existing journals are deleted from the repository. Therefore, rights information from DOAJ need to be updated regularly. Note: The oai-pmh repository does not maintain information about deletions.
2. By definition, DOAJ does not list journals which use embargo periods (e.g. many Highwire journals) or which only provide parts of their content under oa condition (e.g. some BMC journals or backfiles with costs?).
3. Therefore: DOAJ can be used to check if an journal is "on the golden road to OA". According to the DOAJ definition, this information could be escalated to all articles published in the journal. To avoid continuous updates, the information may rather be fetched dynamically than physically stored in pubman. If no information is available, this does not necessarily mean that the journal does not provide OA articles.

Discussion on Journal Abbreviations

Sabine and Traugott:

We do not see a need for tagging the acronyms of e.g. journal titles. The functionality that if an depositor fills in the acronym of a journal title and the full title should then be filled in automatically by the system can - in our opinion - also be provided if the acronym of the journal is stored in the title or alternative title element. The only scenario when tagging might be required is if we would like to generate lists of e.g. journal title acronyms but this can also be done by using "Woerterbuecher" etc.

Comment Natasa: Additional vocabularies are again another level of complexity to my understanding (if I understood your message right  :) . My proposal is to see what e.g. ZDB offers /if they offer abbreviations clearly separated from journal names - then we already have sources to do it in our system and clearly "tag" what is abbreviation and what is an alternative name/. [Comment Inga:] I would like to support this idea: Let's get/fetch/use the information if it's available.

[Comment Natasa:] Yes, will be done.

Comment Inga: Another use case is the generation of reference lists following a citation style which "expects" journal abbreviation, e.g. the ACS citation style.

Comment Traugott: this seems very ambitious to me, yet another piece of information we would need to control in an authority list maintained by escidoc, most probably via storing all the data locally. As all authority lists, creating, de-duplicating, including disambiguation information, validating and maintaining it, is a huge job. To my experience, acronyms are much worse than e.g. journal titles or place names. Sometimes they are even part of the official title of the journal.

In case we wouldn't hold a very rich list of different disciplines and journal communities acronyms, quite often several for the same journal, entering an acronym would result in only a wrong full title being displayed or the wrong acronym added to the citation information. For reference lists and citations (acc. to different styles), authors would very much like to see their (community's) own acronym, I assume. In many cases, they might not know which this is, even if we would be able to display several alternatives.

The main problem we see is that it is most likely that for one journal title more than one acronym is in use and that one acronym might be used for more than one journal, hence acronyms are not unambiguous. [Comment Inga:] ... and unfortunately journal titles are not unambiguous as well! This has to be taken into consideration when offering the functionality that the depositor fills in an acronym and the system provides the full title. So the depositor has to check if the right full title has been filled in or he/she has to select the full title out of a list of potential "right" full titles.

Comment Natasa: To my understanding of course this should be the case. The metadata in the item are populated with journal name and not with the journal abbreviation. When user tries to enter part of journal abbreviation or journal name the system should offer a list of names of journals (i.e. list of journals which have this exact or similar abbreviation or part of the journal name) - only from this list the user selects the appropriare journal name and the metadata are filled in correctly.

Would that be fine?

Comment Inga: Yes, this is required! If a lookup returns more than one journal object, the user needs to explicitly select one journal. In this case, the publisher information is probably a necessary piece of information. [Comment Natasa]: OK, agreed then!

Background information

RDF schema: http://schemas.library.nhs.uk/ApplicationProfile/Journal.rdf

This looks quite comprehensive and we just need a small subset . After 10 minutes analyzing the schema, I'm not sure how the identifiers are further encoded (ISSNURL?). My vote: too complex, reduce it to minimum? --Inga 16:47, 29 November 2007 (CET)

NLM DTD: http://dtd.nlm.nih.gov/publishing/tag-library/2.3/n-z4u0.html

Journal identifier used by CONE

Status: The current implementation does not assume manual additions of journal records directly via CONE, but an automatic ingestion from the SFX knowledge base only. Therefore, the SFX-ID has been used as primary identifier by CONE so far, but this proves to be no reliable solution, e.b. because SFX-IDs are maintained by ExLibris' team only (-> Not all journals will have an SFX-ID already - because they are too new, etc.)

Discussion:

An ID other than SFX should be used as unique ID, as it is not possible to find out SFX-IDs for journal names, which one wants to add. Maybe we can use own IDs like for persons or the ZDB-ID? --Nicole and Karin

I agree, SFX-ID should not be used. We chose SFX because in our initial datasets this was the identifier that was best populated, compared to ISSN. But what else? If we took our own ID we would not have "Control of Named Entities" anymore, just "Named Entities". --MFranke

Please note that the use cases below assume that the "authority" is represented by a special set of pubman users - which could be an argument for following Natasa's proposal below (local identifiers, records extended by authoritative data) --Inga 16:16, 14 April 2009 (UTC)

Please consider that NIMS has high interest in contributing to quality to the journal data, but they do not have any SFX service they could use. I am therefore not sure, if it is good idea to bind the "quality management", i.e the authorisation of entries, to SFX. Still, to offer SFX as optional "quality check" might be of good use at least in MPG. --Ulla

Can we use ZDB-ID instead maybe? This is at least easier to find out than SFX :-). Inga, is ZDB-ID unique? We only have to think about what to do if there is not ZDB-ID. --Nicole 15:41, 14 April 2009 (UTC)

I guess you may run into similar problems with the ZDB-ID as well, i.e. the registry is controlled by an external authority and new journals may not necessarily be available in the moment they are required. I'm not sure how well Japanese titles are covered as well. But, the ZDB-ID is available in a range of systems (incl. EZB) - which may be a big pro.
Additional note: If it's only a "usability problem" with retrieving the SFX-ID: I could add it to the sfx menu as an intermediate solution, e.g. example - or a bit more hidden ;) --Inga 17:08, 14 April 2009 (UTC)

Plans after R5

Proposal by Natasa

1. each new Journal entry, if not selected from a list has a growing (local Id)
2. authorizing it may mean some integration with SFX/ZDB, e.g. by obtaining an SFX-ID via SFX API or ZDB z39.50 interface
   • authorizing it may mean actually even relating it to an existing SFX-ID in Cone (e.g. name alternatives )?
3. the identifier metadata on the publication will not be changed, some interfaces changes though would be required for Cone service.

Another alternative, which would be a bit more heavy would be to:

1. allow only authorized journal entries in the metadata
2. have "journal entry user" who will take requests for new journal entries
3. enter somehow in SFX-ID (that is if any SFX integration would be possible)
4. Only afterwards modify the submission (or enable some automated utility for update of journal identifiers)

Uniqueness of SFX-ID

SFX-ID is not unique as you can see here

• http://dev-pubman.mpdl.mpg.de:8080/cone/rdf/journals/urn:sfx:110975506069213_1 (Zoology - Analysis of Complex Systems, 0944-2006)
• http://dev-pubman.mpdl.mpg.de:8080/cone/rdf/journals/urn:sfx:110975506069213_2 (Zoologischer Anzeiger, 0044-5231)

Ya, I do remember that Natasa informed me about this occurrences long time ago, but I never found the time to follow-up in detail. After checking the incident today, I believe that my lookup-up script may have been confused by SFX-IDs of "relatedObjects" (a feature which have been introduced to the SFX KB in beginning of 2008). For the example above, following SFX-IDs would be correct

• Zoology - Analysis of Complex Systems, 0944-2006: 954925572911, see request at SFX-API
• Zoologischer Anzeiger, 0044-5231: 954925454928 request at SFX-API

... and both objects are related to "Verhandlungen der deutschen Zoologen" with SFX-ID 110975506069213. Sorry! --Inga 15:10, 14 April 2009 (UTC)

How to retrieve an SFX-ID for a journal?

Technically it is already possible to add new entries. To retrieve the SFX-ID you may either

• search for the journal title via MPG/SFX citation linker. The source code of the corresponding SFX menu includes the object-id, e.g.

'rft.object_id' => '954925427230'

• request the SFX API. The context object hash includes the element

<item key="rft.object_id">954925578060</item>

Important Notice: several object identifiers may be listed for one request due to "relatedObject" feature (see examples above)

Potential external sources

The tables give an overview of potential sources of controlled named entities which are of interest. The information given in the tables reflects the current situation and has to be updated from time to time. The tables are in stage "work in progress" and other sources might be added.

Journal

Name of service	Scope	Info	Formats supported	Interfaces	Costs	Access
Zeitschriftendatenbank (ZDB)	ca. 1,3 mio records	Introduction	MAB2, UNIMARC, SUTRS	Z39.50		It has to be clarified with the GWDG if a tailored version of the ZDB (only listing MPG licensed journals) is available.
ISSN Register	1.284.413 records (2006)	http://www.issn.org	MARC21, UNIMARC	Z39.50	costs	Access via the ISSN portal or Z39.50 or via a combined web access Z39.50 and ISSN portal

Rights

Name of service	Scope	Info	Formats supported	Interfaces	Costs	Access
SHERPA/RoMEO Publishers copyright policies&self-archiving	340 publishers (July 2007)	http://www.sherpa.ac.uk/romeo.php	XML	Prototype API	Conditions of re-use	Prototype API
Directory of Open Access Journals (DOAJ)	2.987 journals, 164.284 articles (5th of December 2007)	http://www.doaj.org	XML	OAI-PMH	Conditions of re-use	OAI-PMH

International Standard Serial Number

Elektronische Zeitschriftenbibliothek

Identifier

Zeitschriftendatenbank

Uniform Resource Locator

Open Access

Central European Time

Resource Description Framework

Document type definition

Control of Named Entities

OpenURL Link Resolver

Coordinated Universal Time

National Institute for Materials Science Japan

Application Programming Interface

Koninklijke Bibliotheek

Gesellschaft für Wissenschaftliche Datenverarbeitung Göttingen

Max-Planck-Gesellschaft

Securing a Hybrid Environment for Research Preservation and Access

Extensible Markup Language

Open Archives Initiative Protocol for Metadata Harvesting