Strategies and technologies of sharing in contributor

While we argue about and discuss the plusses and minuses of contributor-run archives, groups formed by people of shared interests and of varied technical competencies have been creating, maintaining, sustaining, and growing their archives for over a decade in several cases.

These contributor-run archives make use of powerful open technologies to facilitate their projects. In this article I will focus on three different volunteer-run projects that involve worldwide cooperation using advanced technologies to further their ends. The Linux Documentation Project, the Degree Confluence Project, and Etree.org are all large projects that involve many contributors with technical teams of various sizes using a variety of technologies. Each project will be described in terms of its aims; its history; its rules, or lack thereof, for contribution; its technologies; and its current state of practice. From these examples we can draw some lessons as well as some enhanced awareness of technologies of cooperation. Among the technologies used by the projects are wiki, mailman, Shorten (SHN), FLAC, PHP, mySQL, PHPbb, Postnuke, BitTorrent, rsync, XML, and CVS. All of these technologies are "open" and available for installation, customization, and further sharing of their code.

Over my dozen years as director of ibiblio.org and its predecessors, sunsite.unc.edu and metalab.unc.edu, I have seen many projects flourish and many projects stagnate and more than a few projects die completely. At the time of this writing, ibiblio.org hosts and facilitates over 1,500 projects in addition to our extensive software collections. In May of 2001 I published a brief article describing how open-source tools might be used in contributor-run libraries (Jones, 2001, pp. 45-46). For this article I aim to describe some successful examples of open-source contributor-run collections. I have selected three projects with worldwide contributor bases, innovative technology use, open management, and volunteer staff for consideration. All three projects solicit participation from their users, and amazingly they have consistently received reliable and enthusiastic contributions. As a result, each is--thin its particular area--a must-visit resource.

Briefly, the Linux Documentation Project aims to provide reliable, accurate, and helpful documentation to Linux users from beginners to advanced systems administrators in every language in the world. The Degree Confluence Project aims to document the world by visiting every degree confluence on the earth. A degree confluence is defined as "the exact spot where an integer degree of latitude and an integer degree of longitude meet." Confluence.org volunteers participate in creating a database of photographs and narrative descriptions of their visits to each degree confluence on, or near, dry land on the entire earth. Etree.org aims to provide a forum for the exchange of very high-quality concert recordings of"tape friendly" bands (Etree.org, 2004e).

Each of these projects has an education component as part of its mission-that is to say, guidelines and FAQs for new contributors and new users. These educational components also serve to advance the ideology of sharing the information, skills, and experience that is a part of each project.

THE LINUX DOCUMENTATION PROJECT

The Linux Documentation Project (2004), begun by Matt Welsh in 1992 not long after the first wide release of Linux itself, predates the World Wide Web (Garrels, 2004; M. Garrels, personal communication, May 14, 2004). The goal of the project, as described by volunteer David Lawyer in the Linux Documentation Manifesto, is "to create the canonical set of free Linux documentation. While online (and downloadable) documentation can be frequently updated in order to stay on top of the many changes in the Linux world, we also like to see the same docs included on CDs and printed in books" (Lawyer, 2000). Thus, while the Linux Documentation Project can be seen as a long-lived online community project, its goals are not limited to cyberspace; the project aims for world conquest--or at least to conquer the world of Linux Documentation. To a large extent, TLDP--as it is now known--has succeeded. Andy Oram of O'Reilly and Associates, a leading technology publishing company that might be considered to be the competitor of TLDP, has written that TLDP "is an impressive organization that has editors, guidelines for reviewers, procedures for updating documents, translators--in short, it's an organization that has tried to reproduce everything about conventional publishers, but in an open and volunteer manner" (Oram, 2004). Oram also praises TLDP as "a phenomenon we should all be following as a model for documentation in an open source community" (Oram, 2004). I would note that an organization of a dozen years is no longer a phenomenon but is, in the world of cyberspace, an institution.

Those wishing to contribute documents to TLDP are pointed to a detailed yet straightforward Author's Guide that describes what and how to participate. The process goes as one might expect from any publishing company:

1. Become familiar with the Linux Documentation Project's other works by looking over the site and joining the Discuss mailing list.

2. After having identified a gap in the documents or that a new document is needed, propose your document to the Discuss mailing list, including if possible an outline and description of the document.

3. Write your document.

4. Mark up your document or seek help in doing mark up. All documents published by TLDP are in SGML, Docbook XML, or LinuxDoc formats to allow for flexible republication. Obviously, this might constitute a high barrier of entry for contributing writers, but TLDP volunteers have agreed to work with new contributors by instructing or even providing proper markup for submitted documents (Sundaram, 2003).

5. Submit your document for review by sending a copy or a link to a copy to the Submit mailing list. A language editor, a technical editor, and a metadata editor review all documents. It is not unusual for all three editors to actually be the same person. This process could take up to two weeks. Of particular note to readers of Library Trends is the requirement that eleven metadata fields be complete and accurate before a document is accepted. Metadata editor Emma Jane Hogbin writes that the goal is for TLDP documents to be Open Archive Initiative-compliant within the year (E.J. Hogbin, personal communication, May 12, 2004).

6. After, or even during, the review process, the document is added to the Concurrent Version System (CVS) for TLDP. While the use of CVS is optional, it is a great innovation. CVS allows an author to keep an offsite copy of the document as well as allowing other authors or editors to make traceable changes in the document. Additionally, the change log may be included in the document automatically as an aid to readers. Ideally in the future, the change log will also interact with the appropriate metadata elements and will be used to announce the new or newly revised document to the TLDP Web site and appropriate lists (E. J. Hogbin, personal communication, May 12, 2004).

Other than mailing lists and CVS, TLDP's use of technologies of cooperation is minimal but highly effective. Requiring metadata makes documents easier to find and to use in a trusted manner. Choosing an open markup language, XML, and an open tag set, Docbook, as opposed to Word or even PDF, gives the documents flexibility and longevity. That flexibility includes the ability to use open-source publishing tools such as openjade, the dsssl stylesheets, libxml2, xsltproc, XSL, etc. (Hogbin, Komarinski, Godoy, & Merrill, 2004).

Because of the strong and mostly friendly involvement of the editors, there are few documents rejected. In fact, most rejections occur at the proposal stage and usually because of content overlap with an existing document. Even then the proposing author is encouraged to contribute to the overlapping document and so to be a part of the continuing process (G. Ferguson, personal communication, May 12, 2004).

The current TLDP archives hold works from over 500 different authors, although not all of the authors are active at one time. There are 345 subscribers to the discuss@en.tldp.org list, 52 active editors, and a core team of 19, as well as translation coordinators working in languages from Albanian to Walloon. The impact of TLDP cannot be overestimated. Not only are TLDP documents distributed with major Linux software distributions, but also the entire site is mirrored or copied completely on over 300 official sites around the world (see http://tille.soti.org/images/tldp-world.jpg).