WikiReference

WikiReference: A Proposal for a Wikimedia Indexing Project

I would like to propose a new Wikimedia project that I am calling WikiReference.

The basic idea is simple: WikiReference would be a free, collaboratively created index of published material. In a sense, it would be to indexing what Wikisource is to source texts.

Wikimedia has Wikipedia, which tells us what is known about a subject. It has Wikidata, which provides structured information about people, places, works, and other entities. It has Wikisource, which makes many historical and public-domain texts available. What it does not really have is a place where people can collaboratively build an enormous index showing where subjects have been discussed in the published literature. This proposal is not intended simply to duplicate Wikidata. Wikidata can store bibliographic information, but the proposed project would provide a user-facing environment specifically designed for the collaborative indexing of publications and their contents, including periodical-by-periodical indexing, issue-level organization, subject indexing, and discovery of literature that may not otherwise have a structured presence in Wikimedia projects. The question I am asking is whether this functionality should exist as a dedicated Wikimedia project or be developed within existing Wikimedia infrastructure.

That is the gap I think WikiReference could fill.

Take a simple example from my own field of music. Suppose someone wants to research Ralph Vaughan Williams. Wikipedia can provide an overview of his life and work. Wikidata can provide structured information about him. But what if the question is: Where has Vaughan Williams been discussed in the periodical literature?

The answer is scattered across hundreds of journals, magazines, newspapers, reviews, and other publications. Some of it is in commercial databases, some in library catalogs, some in specialized bibliographies, and some is "darker" data that is not indexed at all. The contents of many periodicals are known only from consulting hard copies that collect dust in various physical locations.

WikiReference could bring this material together.

A journal such as Music & Letters could have a page for the journal, pages for its volumes and issues, and individual entries for the articles and reviews in those issues. An article could be connected to its author, its subjects, the people it discusses, and the musical works it discusses. Just the metadata, to avoid copyright entanglements.

The same system could work for newspapers, magazines, scholarly journals, conference proceedings, book chapters, exhibition catalogs, newsletters, trade publications, and eventually almost any kind of published material.

The important thing is that WikiReference would not be another Wikipedia. Wikipedia tries to explain a subject. WikiReference would answer a different question:

Where has this subject been discussed?

If I searched for Vaughan Williams, for example, I would want to be able to see not only an article about him, but an interconnected record of the literature about him: articles, reviews, books, chapters, newspaper pieces, and other publications, ideally arranged chronologically and linked to the publications in which they appeared.

This could be particularly valuable for historical research. Imagine being able to search for a composer, politician, artist, historical event, scientific subject, or place and immediately see the published literature associated with it, including obscure and long-defunct periodicals that are not well represented in the major commercial databases.

The project could also work particularly well with existing Wikimedia projects. A WikiReference entry could link to Wikidata records for authors, journals, people, places, and works. Where a publication has been digitized in Wikisource, the index could link directly to the text. Wikimedia Commons could provide scans or images where appropriate.

I think the most important aspect of the project is that it should be possible for ordinary people to contribute without having technical knowledge. Someone who knows a great deal about nineteenth-century newspapers, British music, genealogy, baseball, local history, or any other subject should not have to know how to program or build a database in order to create a useful index.

That is where I think the Wikimedia model could be particularly powerful. No single institution could possibly index everything. But thousands of people with specialized interests could collectively create an enormous resource.

I have actually tried to develop something along these lines myself. I wanted to build a large periodical index, but I quickly discovered that the technical requirements were much greater than I could handle, and the student programmers who were helping me were not equipped to build the necessary infrastructure. That experience convinced me that the underlying idea is more important than any particular implementation.

I am therefore not proposing that Wikimedia simply build a website for me. I am proposing that the Wikimedia community consider whether a collaborative indexing project would fill a genuine gap in the existing ecosystem.

A small pilot would be enough to demonstrate the concept. One could begin with a handful of periodicals, index their contents, connect the entries to Wikidata, and see what becomes possible when the literature is treated as a network rather than as isolated bibliographic records.

The larger vision would be a free World Index of published knowledge—something that could eventually encompass scholarly literature, newspapers, magazines, books, reviews, historical publications, and many other forms of printed material.

In the simplest terms, I think of it this way:

Wikipedia tells us what we know. Wikisource gives us the sources. Wikidata tells us what the things are and how they relate. WikiReference tells us where the published discussion can be found.

I think that could be a very useful addition to the Wikimedia ecosystem.

Proposed by

Ryan Ross, RMR1108