The Distributed Digital Library of Chant Manuscript Images (DLL) is an annotated index of images on the WorldwideWeb, maintained by the NEUMES Project. The manuscript images in this 'virtual library' are physically located at many unrelated sites on the Web. Generally, we refer the user directly to the image-hosting site rather than hosting a copy of the image ourselves. The main exception to this rule is that, we may provide a small excerpt of the image just to assist the user to visually identify the manuscript, or to highlight an interesting feature of the manuscript that may motivate the user to visit the hosting website and examine the image in detail.

We provide a 'value added' to the images by our annotations, proprietary indexing scheme, and coverage of manuscripts across many websites. This service has social value as follows.

1. In one step, users can search across many websites for images of medieval manuscripts that contain particular chants (such as, chants for the Matins Office of Trinity Sunday in the Roman liturgy) or manuscripts written in a particular geographical region, historical period, or musical notation (such as, examples of early Aquitanian notation), without having to search repetitively at each separate website. Users might not know a priori which manuscripts contain the desired content, nor which physical library houses a manuscript. Also, some manuscript-image hosting websites might be unkown to the user.

2. The user's search effort can be more productive, because just one search grammar must be mastered. By contrast, each image-hosting site may have its own, special grammar of searches, which must be mastered by the user. A search also can give more consistent, complete results due to the unified scheme of metadata by which this service catalogues all manuscript sources.

3. The user can search for manuscript images that are hosted by websites that do not have their own search capabilities, or that do not provide descriptive metadata with the images.

The Distributed Digital Library of Chant Manuscript Images is a service provided by the NEUMES Project for non-commercial, educational, and cultural purposes. It is not published in print.

---

The images indexed by the DLL are made available online by universities, public libraries, monasteries, private collections, and so on—hence the concept of a distributed digital library. The DDL is a selective index of Web content—that is to say, it is a searchable guide covering a coherent topic. It has an idiomatic perspective by which it organises and presents information; this includes commentary or annotation on individual items in this 'virtual' collection. The DLL is just one of many possible indexes that would cover the same content; each such index can be specialised to the interests of a particular user community. The overarching purpose of the DLL is to assist scholars in finding images online of primary-source manuscripts containing neume notation.

Some image-hosting sites provide comprehensive information with each of their images, while other hosts give little or no explanation about manuscripts. Across hosts, the kinds of metadata provided, the manner in which dates are written, keywords used, and so forth, may be inconsistent; such inconsistency makes it difficult for users to find manuscript images by consistent search methods. Of course, an image by itself is not ordinarily searchable for text or musical content of the imaged manuscript. (By contrast, manuscript texts that have been transcribed to 'character data' in HTML, XML, Word, PDF, etc., are searchable.) Furthermore, much of the important information about a manuscript is not found in the manuscript.

The DDL not only provides hyperlinks to online images of chant manuscripts; it also maintains its own metadata about each imaged manuscript (including the species of neume notation, geographical and religious congregation of origin, estimated date, and so forth). The user can search these metadata in a consistent way via the DLL's full-text query capability. Maps, historical timeline, etc., are provided to assist novices in their understanding of the materials and to enable alternative methods of querying the sources. Showcase Galleries also are provided so that users can browse a small collection of special-interest sources. The Result Set from a user query includes link(s) to the corresponding NEUMES transcription(s) if any exist.

Images in the 'virtual' library of the DLL have been selected 'by-hand'. Metadata about the manuscripts also have been scholarly-edited. An important focus of the DLL editorial process is to illustrate the many kinds of neume notation that were used for writing chant, spanning more than a thousand years of history and geographical area including all of Europe, parts of the Middle East, and Egypt.

The concept of a 'distributed digital library' is intuitively understandable to most users of the WorldwideWeb. It is a 'virtual' library whose content is the aggregation of many independent websites, rather than a physical collection that is kept in a central repository. This architecture is helpful in the environment of the Web for several reasons: the work of creating digital images of manuscripts—and maintaining websites for them—is distributed among many institutions; the rights of the institutions that own the manuscripts are inherently protected; sustainability of the DLL in the long term is more feasible because of decentralization; the specific needs of a particular user community can be met without compromising the generality of hosting websites; and so on. In Figure 1, below, we call each such website a 'resource'. Typically, the image content at a resource is unique to that host.

The architectural model of the DLL is illustrated in Figure 1 in terms of a data-flow diagram.

Diagram Key :
	a domain (or, address space) on the Internet
	a server accessible on the Worldwide Web
	a digital document, such as an image of an artefact, NEUMES transcription, chant metadata in XML, essay in PDF, or HTML
	a reference (or, 'pointer') to a resource document
	view a document

Data Flow :
The Digital Library Index provides a list of resource documents, from which the end-user can make a selection; this happens in a dialog between end-user and Index server. When the end-user has selected a document of interest, the reference to this document is passed from the Index server to the user's Web browser. The user's browser then downloads the referred-to content  directly from a resource server for viewing.

Figure 1. Architectural model for client/server interaction of the Distributed Digital Library of Chant Manuscript Images.

• See also: Barton, Terzopoulos, and Craig-McFeely, Online Access to Manuscripts of Byzantine Chant," §3 "Digital Libraries," for a more detailed discussion of this architecture. [Format: PDF. Filesize: 4 MB.]

The software infrastructure of the DLL is designed to make this resource inherently self-sustaining via voluntary, collaborative contributions of content by the community end-users. Contributions ordinarily shall include citation of the URL [Uniform Resource Locator, or Web-address] where a manuscript image is hosted, and submission or redaction of metadata describing the manuscript and its content. The metadata—in the format specified by the DLL—may be paraphrased from metadata at the image-hosting site or from other references, facts that don't involve original authorship (such as the dimensions of the manuscript), and/or a description authored by the person making the submission.

The goals and architecture of the DLL are significantly different to the Wikipedia online encyclopedia. Useful lessons can, however, be gleaned from the Wikipedia experience as a paradigm for collaboratively-edited, online reference resources. In a seminar talk [excerpted by the NEUMES Project, below], Jimmy Wales (principal of Wikipedia) discusses some points that are relevant to the DLL. The seminar was moderated by Stewart Brand (founder of the Whole Earth Catalog). Four of Wales' points are especially relevant.

• He sees a powerful trend in society: that knowledge and culture shall be freely available to everyone via the WorldwideWeb. This trend contrasts to the era of printed books and journals, when the high costs of printing and distribution justified making people pay for access to information. He challenges the theory that, in order for people to produce high-quality work, they must be compensated financially. Wikipedia, he says, demonstrates that people will contribute their time to a collaborative effort, if this work is likely to have lasting value for the community.
• He challenges the idea that only persons with sufficient academic credentials or members of a scholarly group are able to make reliable contributions to the body of human knowledge. He argues that many non-academicians have valuable contributions to make, and that peer review is sufficient for ensuring the reliability of content on Wikipedia.
• He finds that, in the various language-branches of Wikipedia (as with open source software development projects), it is necessary to have a "benevolent dictator" who controls a project during its early stages. He says that a core group of editors (he mentions the number twenty) in each specific area of interest or expertise can eventually assume this role.
• He says that Wikipedia is hiring professional programmers to create software infrastructure for Wikipedia. In the past, he says, Wikipedia. has relied on volunteers to create its software. He says, however, that relying on volunteers for programming (as distinct from volunteers for content creation) has not been a successful strategy for Wikipedia.

Jimmy Wales, "Vision: Wikipedia and the Future of Free Culture"; speech presented at the Seminars About Long-term Thinking; San Francisco, California; 14 April 2006. [Length, excerpts: 23 minutes. Format: MP3 streaming audio. Filesize: 4 MB. JavaScript required.] The full speech [length: 1 hour, 16 minutes] is available at The Long Now Foundation.

The requirements of the DLL differ from Wikipedia in two main ways, as follows.

1. The DLL is specialized for just one field of interest: manuscripts that contain medieval neume notation. Wikipedia accommodates information from all fields of interest. The specialization of the DLL allows the structure (or, ontology) of its metadata to be tailored just to this one field. Search queries, search controls, content-browsing, and so on, likewise can be specially adapted for this field. DLL searches can be made to appear more 'intelligent', because a significant amount of expert knowledge about the field can be engineered into the software infrastructure and into the metadata ontology.

2. Wikipedia's strategy is to allow entries, or modifications of entries, to be made by any end-user. The DLL has an Editorial Panel of experts in the field of chant musicology. New data entries or data modifications coming from the user community shall be approved by a member of this Panel before being incorporated into the DLL. The Editorial Panel, however, is not meant to be a static group of 'insiders': it shall be self-sustaining over time by the addition of new members. The process for adding new members to the Panel shall be that: a new Editor shall be added to the Panel by approval of two or more existing members of the Panel. A 'veto' power shall reside in the System Administrator (typically, a non-musicologist) solely for the purpose of preventing a malicious 'hijacking' of the Editorial Panel by an inappropriate group. In the field of chant musicology, this divergence from the Wikipedia strategy is deemed (in the opinions of many chant scholars) to be highly important for maintaining the scholarly reliability of the DLL content.

If nothing else, Wikipedia at least has occasioned serious discussion on whether systems of voluntary contribution of content will—through community dialog—eventually produce reliable reference resources. The main argument for voluntary contributions is that the contributors are not paid; therefore, due to the Internet, the work product can be distributed for-free to anyone (the 'free culture' argument). The arguments for centralized control by an editorial panel point to the risks of malicious vandalism, self-interest or bias, and inaccuracy of contributions. The Wikipedia experiment suggests that these problems can be solved by internal policing by the community. What remains an open question, however, is whether a 'group think' mentality inevitably infects a community of self-appointed, non-expert contributors, such that the community's viewpoint necessarily is nonobjective.

A negative view of Wikipedia is expressed by Dr David Hill (World Innovation Foundation), who writes: "The greatest problem with Wikipedia that we now find is that they are highly selective in who should place information and where, therefore they will never really have a web-based encyclopaedia that is unbiased and totally factual. It is totally at the whims of the few enlightened ones who control what should be a great reference" [cf., The Long Now's "Blog Summary" of Wales' speech].

Jimmy Wales at his Macintosh computer.

Wales has been taken to-task for editing his own biography in Wikipedia: "Having seen edits like this," says co-founder Larry Sanger, "it does seem that Jimmy is attempting to rewrite history" [cf., Wikipedia, s.v. "Jimmy Wales"]. "Even Wales has been caught airbrushing his Wikipedia entry—eighteen times in the past year. He is particularly sensitive about references to the porn traffic on his Web portal" [New Yorker Magazine, 31 July 2006].

The Boston Globe reports that Wikipedia's policy of allowing anyone to edit its entries has led to "thousands of flattering or disparaging changes in profiles of dozens of members of Congress" [David Mehegan, "Bias, sabotage haunt Wikipedia's free world," Boston Gobe, 12 February 2006]. "Wikipedia touts itself as 'the free encyclopedia that anyone can edit,' and it is exactly that quality that is causing problems," the article continues. "[T]he dark side of that freedom is that Wikipedia's articles are becoming battlegrounds, pitting writers with biased viewpoints and vandals trying to sabotage entries ...." Wales favors the current policy of 'open editing'; Sanger (who has now left Wikipedia) believes that, "supervision should be in the hands of specialists," says the Globe.

Gregory Fried (Chair of the Philosophy Department at Suffolk University) comments on Wikipedia: "I don't doubt that it has good articles, but I don't know which are good and which are not" [ibid.]. Some university professors complain of students simply quoting Wikipedia,—without citing any reference—in their 'research' papers in lieu of due diligence on homework assignments. This is not, however, the fault of Wikipedia: there were ways to short-shrift one's homework long before the advent of the Web. It would seem incumbent on educators at some level to explain to students the difference between primary and secondary sources, the distinction between scholarly-edited and popular articles, and the need for using quotation marks and citing one's sources in any case.

By contrast, the prestigious journal Nature finds that, "the accuracy of science in Wikipedia is surprisingly good: the number of errors in a typical Wikipedia science article is not substantially more than in Encyclopaedia Britannica, often considered the gold-standard entry-level reference work" [Jim Giles, "Special Report Internet encyclopaedias go head to head," Nature, 438 (15 December 2005), pp. 900-901]. Britannica has disputed this claim [Encyclopaedia Britannica, Inc., "Fatally Flawed; Refuting the recent study on encyclopedic accuracy by the journal Nature," March 2006]. Nature, in turn, rebutted Britannica in a statement of 23 March 2006. The crux of the disagreement seems to be that, in some cases, Nature compared Wikipedia articles to articles in Britannica's Book of the Year, Student Encyclopedia, or website rather than comparing to the Encyclopaedia Britannica, proper.

Some scholars have a low opiinion of the Worldwide Web because of the great amount of unreliable and inaccurate content they find when they search in Google or other full-text indexes that try to cover the full extent of the Web. Such censure seems misdirected: it impugns free speech on the Web, when it would more appropriately be directed at indiscriminate indexing methods of robots. In the landscape of dubious quality on the Web, Wikipedia seems relatively dependable for accuracy—and it is free.