The Text Encoding Initiative's represent a significant advance in the effort to make computing machinery useful in scholarly research. They provide a well defined method for the electronic representation of textual material, which is independent of any one software or hardware system, suitable for use with texts in any natural language, and designed to meet the requirements of scholarship, rather than to assure the convenience or profits of software vendors.

In defining a basic vocabulary for the electronic markup of textual objects, the Guidelines constitute a preliminary census of the types of textual objects researchers most often recognize and work with in practice. The Guidelines thus constitute a major contribution to the theory and practice of computational work in the humanities. Further work can build on the foundation already laid to extend the coverage of the Guidelines to provide better methods for encoding some types of texts and some types of information about texts.

This project aims at assuring the adequate dissemination of the Guidelines to practicing scholars in the field, by means of additional user-level documentation, books of examples, tutorials and workshops, and consultation with humanities projects actively creating electronic texts. It thus provides a means of making good on the promise of the Guidelines and ensuring that humanistic research can capitalize on the opportunity they offer, that electronic resources once created, sometimes at great cost, can be used and re-used in many different ways; that electronic texts can survive longer than the short-lived systems we use to create them; and that the work done by today's scholars can still be utilized by those yet to come. Narrative Description Introduction: Nature and Significance of This Project

Computers are revolutionizing scholarly research with texts, as well as many other parts of our culture. With the aid of computers, texts may be created, stored, manipulated, edited, and distributed much more quickly and reliably --- and in the long run, much more cheaply --- than is now the case with non-computer techniques. The increasing use of computers to prepare and manipulate texts, however, leads to an increasingly acute need for useful methods of representing texts in electronic form, with all their intrinsic variety and complexity. Such methods have been an object of interest since humanists began, in the late 1940s, to use card-punch machines to encode texts in machine-readable form. But not until recently has there been a serious effort to develop a general-purpose encoding scheme adequate to the needs of scholarly research with textual materials. The Need for Text Encoding Guidelines

Over the decades, scores --- probably hundreds --- of schemes have been developed for encoding in machine-readable form the characters of a text, its logical divisions, its annotations and other apparatus, and information relevant to its processing for a particular purpose (such as morphological or semantic information). Such encoding allows us to mark explicitly features of the text which are otherwise only implicit, in order to allow them to be analysed automatically. Many of these schemes were designed from scratch as input formats for specific analytical software packages (concordance programs, content analysis software, parsers, formatters, word processors, etc.), while others were developed to suit the needs of specific large general-purpose text corpora. Many are woefully inadequate for scholarly work with texts, and even those designed with research work in mind often focus on rather specialized areas and do not suit the needs of other fields. Moreover, since most existing encoding schemes were developed on ad hoc principles, conversion between different schemes remains difficult, even when both the source encoding scheme and the target scheme are both well documented, which is unfortunately not always the case.

The multiplication of such incompatible text formats wastes time, effort, and money. Because no single format is generally accepted, scholars creating new electronic texts must create their own new format, or else choose one from among the many incompatible existing formats. If, as often happens, an existing scheme reflects only the research interests of its originators and the peculiarities of the texts they studied, it may be so difficult to generalize the scheme to other texts or other interests that it would be easier to develop an entirely new scheme, despite the cost in time and effort. Would-be users of archived or recycled texts thus often must choose between deciphering an existing encoding from scratch or transcribing the text again in their own format. In either case, the process involves a wasteful duplication of effort and a distraction from the main purpose of research.

Time and effort are also wasted by the inadequacy of many commonly available text formats for scholarly research. Because common text processing software is not designed for scholarly applications, common text formats often lack even the most rudimentary facilities for encoding textual variants, multi-level annotation, synchronization points between parallel texts, etc. Researchers who use these common formats must expend substantial effort on attempting to circumvent the limitations of their software and text format.

A common text encoding scheme developed for the needs of scholarly research would minimize these problems. Scholars would not have to develop personal encoding schemes from scratch; documentation of encoded texts would be simpler; duplication of effort would be reduced. Such a common scheme can also simplify the creation of software for academic or commercial purposes, by providing a target text format for the software to support. The Text Encoding Initiative

Arguments like those outlined above led, in 1987, to the creation of the Text Encoding Initiative (TEI), an international cooperative effort (funded in part by the National Endowment for the Humanities) to develop and disseminate guidelines for the encoding and interchange of machine-readable (i.e. electronic) text for research purposes.

With the participation of scores of practicing researchers on its working committees and work groups, the TEI has, since 1988, systematically developed a new encoding scheme for electronic texts, with the needs of researchers explicitly in mind. The TEI Guidelines have been designed to: suffice to represent the textual features needed for research be simple, clear, and concrete be usable without special-purpose software allow the rigorous definition and efficient processing of texts provide for user-defined extensions to the scheme conform to relevant existing and emergent standards

A first public draft of the Guidelines was published in July, 1990, for public response and comment. A revision taking into account the reactions to the first draft began appearing chapter by chapter in 1992, and the Guidelines proper are expected to be published formally by the end of 1993. (For further details on the history of the TEI, see section , below.)

With the publication of the TEI Guidelines, the research community will have at its disposition a scheme for representing text in electronic form which: can be used in interchanging documents among unlike machines, software, or projects can equally well be used for the creation of new electronic texts is at least as expressive as all major pre-existing encoding schemes for electronic texts, and thus allows any existing electronic text to be translated into a TEI-conformant form without serious information loss provides a flexible set of guidelines adaptable to the different needs of different researchers and projects, rather than a single set of rigid requirements allows scholars to record as much or as little information about the text as they wish allows for user-defined extensions to the encoding scheme, to allow those who work with unusual texts, or who wish to encode categories of structural or analytic information not included in the scheme, to represent that information in their text does not depend on a single hardware or software platform can be used for texts in any natural language is not biased towards any one type of processing or application, but can be used to encode texts for any type of processing desired allows text, images, and other types of material to be combined into complex information objects for multimedia or other presentation defines methods for documenting electronic materials for purposes of bibliographic control and intellectual clarity in a TEI header, which can be used to assist in the management of text, images, or electronic material of any kind

The Guidelines define their encoding scheme in terms of the Standard Generalized Markup Language (SGML), an international standard language for defining markup languages: that is, specifications of the markup or tagging defined for, and allowed by, a given encoding scheme. The tags of a markup language provide a fundamental vocabulary for identifying different segments of a text as being textual objects of a particular kind, with particular attributes. (For further non-technical description of the nature of markup languages, see TEI document TEI ED W33, a copy of which is included with this proposal.)

The goals set initially for the TEI have thus been substantially achieved. To ensure that the benefits for scholarship originally foreseen for the project actually accrue, however, we must take steps to capitalize on the work done so far. What is Needed Now

To make good on the promise of the TEI Guidelines, the work of the TEI must be made accessible to practitioners in the field. This means a concerted effort must now be made to disseminate the results of the Initiative to a larger community, and to formulate those results in forms comprehensible and useful to individual scholars and projects.

The TEI Guidelines themselves are formulated as a reference manual for the encoding scheme they define. They provide a full formal definition of the TEI encoding scheme, and document in the necessary detail just what each tag in the markup language means and how it may be used. Since the Guidelines cover such a wide variety of text types and define markup techniques for such a broad range of specialized scholarly uses, they have also grown rather large; as noted above, the published version of TEI P3 is expected to contain about 1200 letter-size pages. Both the level of detail and the size of the document make it difficult to use for those not already fairly expert in the use of computers and the definition of formal languages for data representation. Documentation more accessible to practicing scholars, intensive classes and training workshops, examples of the use of TEI in a variety of scholarly fields, and consulting on difficult problems are all essential if the community is to be able to make full use of the work performed by the TEI's work groups and working committees.

This proposal asks for funds to provide those essentials, by means of four basic areas of activity: development of additional documentation of the TEI encoding scheme, including introductory manuals aimed at practicing scholars in various disciplines a series of workshops and tutorials held both independently and in conjunction with meetings of major professional societies creation of a book of case studies showing the use of the TEI Guidelines in the solution of real research problems consultation with selected projects engaged in the creation of major electronic text resources For details of the work proposed, see section . History and Duration of the Text Encoding Initiative

The TEI has its origin in a conference of experts on electronic text convened at Vassar College in November, 1987, by the Association for Computers and the Humanities (ACH), with the support of the Endowment. The thirty-odd attendees from North America, Europe, Israel, and Japan considered the problems in humanities computing arising from the lack of any commonly accepted practices for text encoding, and agreed that the development of a common encoding scheme could and should be undertaken. At the Poughkeepsie conference, the Association for Computational Linguistics (ACL) and the Association for Literary and Linguistic Computing (ALLC) agreed to join with ACH in sponsoring the effort to develop a text encoding scheme suitable both for the interchange of existing electronic materials and for the creation of new materials.

Following the Poughkeepsie meeting, a Steering Committee of representatives from each sponsoring organization developed an organizational framework and a plan of work for the TEI. The plan foresaw a multi-phase project: initial planning would be followed by drafting; one or more cycles of publication, comment, and revision; review and approval by an Advisory Board representing appropriate learned societies and professional bodies; and finally by a final phase of publication and maintenance of the Guidelines. Structurally, the project would be led by two editors appointed by the Steering Committee, who would oversee and coordinate the work done by a set of working committees and work groups which would be responsible for actually drafting the Guidelines. An Advisory Board made up of representatives from a variety of participating organizations would be responsible for seeing that the interests of various constituencies were adequately addressed, and would be asked, at the conclusion of the work, to approve the results.

Allowing for a certain amount of adjustment to changing circumstances, the project has since gone essentially as planned. Because interest in the project was even greater than expected, however, with concomitant demands for both deeper and broader coverage, the size of the Guidelines has increased several fold; this has delayed the final appearance of the Guidelines.

In 1988, the TEI proper began with a first cycle of development, funded by a two-year NEH grant and a contract with the Commission of the European Communities, and later as well by a grant from the Andrew W. Mellon Foundation. A number of organizations were invited to participate in the project by naming representatives to an Advisory Board (list of members in the appendix), chairs were named for four working committees, and preliminary planning were carried out. In early 1989, the Advisory Board held its first meeting and nominated practicing researchers to serve on the working committees; the working committees themselves began their work in mid-1989. In July, 1990, the first development cycle was completed and its result, a first complete draft of the TEI Guidelines, was issued for public comment as document TEI P1. In a little under three hundred pages, TEI P1 defines SGML tags for basic text structure, common textual features of the sort handled by existing document production systems (quotations, footnotes, bibliographic references, etc.), text-critical apparatus, analysis and interpretation of text (with examples from linguistics), and in-file documentation of the source edition and its treatment in the electronic text.

A second cycle of development followed immediately, in which the coverage of TEI P1 was refined and extended: work groups were formed to address issues of character sets, text criticism, hypertext, formulas and tables, literary prose, verse, drama and other performance texts, general linguistics, spoken texts, literary studies, historical analysis, print dictionaries, computational lexica, and terminological data. In July 1991, the TEI held public workshops in Providence, Rhode Island, and Oxford, England, which introduced the basic principles of SGML and of TEI P1. In early 1992, a revised version of the Guidelines, which was given the document number TEI P2 (for proposal number 2) began to appear, chapter by chapter. In June of 1993, the TEI Advisory Board held its second meeting, at which it reviewed the current draft of both published and unpublished chapters and approved the work done to date and the plans for its completion.

At the time this proposal was being prepared, the published chapters of TEI P2 included: introductory material on the nature of the TEI Guidelines, an introduction to SGML, an overview of the structure of the TEI encoding scheme, a discussion of character representation issues, a definition of the TEI header (which documents the electronic text), a set of tags available in all TEI documents (the core tag set), a tag set for basic text structure (front matter, body, back matter, and major text divisions like parts, chapters, and sections, or act and scene), tag sets for prose, verse, drama, spoken materials, and terminological material, optional additional tag sets for segmentation and alignment of texts (which includes hypertext linking) and for language corpora, and a formal grammar of the TEI subset of SGML. A number of tag sets and chapters on technical topics remain to be published, and are expected to appear during the autumn of 1993. After they have all appeared, some chapters will be revised and the cumulative draft will be printed as document TEI P3. TEI P3 is expected to fill about twelve hundred letter-size pages: a four-fold increase in material since the first public draft three years earlier. Project Staff and TEI Organization

The activities described in this grant proposal will constitute one portion of the overall activity of the Text Encoding Initiative during the grant period. Accordingly, this section describes first the overall organization of the Initiative, and then discusses the allotment of work under this grant. Overall Organization of the TEI

The TEI is sponsored by the ACH, the ACL, and the ALLC; two representatives of each of these three societies make up the Steering Committee, which is responsible for the overall planning and organization of the Initiative. (List of members in the appendix.) The Steering Committee will be responsible for reviewing the work done under this grant, and so funds have been requested to fund one of its meetings during each year of the grant; other meetings will be funded separately.

The broader community is represented in the TEI by an Advisory Board comprising representatives of (at the moment) fifteen professional societies in disciplines concerned with textual study and with electronic texts. Members of the Advisory Board circulate TEI drafts among the membership of their organizations, channel comments back to the TEI, and ensure than the particular needs of their discipline are adequately represented in the project. Several organizations represented on the Advisory Board have expressed an interest in devoting conference sessions at their annual meetings, or pre-conference workshops, to the TEI; their representatives on the Advisory Board will help coordinate the scheduling and organization of these activities. (See also section .)

The technical work of the TEI --- notably the drafting of the Guidelines and the preliminary analyses of problem areas needed before any draft can be prepared --- is performed by a number of working committees and work groups named from practicing researchers in the field in question; the specific organization of the working parties changes over time and is currently under review. Technical work on revisions and extensions of the Guidelines will continue for the foreseeable future; this work, however, is not included in the activities funded by this grant, and no funds for technical work by TEI working parties is requested in this proposal. Members of TEI work groups form a natural pool of experts from among whom at least some of the consultants used in this project can be drawn.

The day to day operation of the TEI is supervised by an editor in chief, C. M. Sperberg-McQueen, of the University of Illinois at Chicago (UIC). With the associate editor, Lou Burnard, of Oxford University, Sperberg-McQueen is responsible for coordinating the work of the committees and work groups and integrating their results into an intellectually coherent whole. At present, Sperberg-McQueen works full-time on the project, and Burnard works three-quarter time on the TEI and one-quarter time on a corpus of contemporary British English. The editors will assume final editorial responsibility for the documentation to be prepared under this grant proposal. Staff Involved in the Dissemination Project

The dissemination work described in this proposal is to be done primarily by the editors, with consultation from the Steering Committee and with the assistance of outside consultants and a project office staff located in Chicago. Specifically, the project will involve work by: the editors of the TEI (C. M. Sperberg-McQueen and Lou Burnard), an office manager / program coordinator, to work in the Chicago office of the TEI a student assistant, to work in the Chicago office of the TEI experts who will assist in the creation of documentation for specific discipline areas, as described below in section (these are listed in the budget as consultants) individuals, usually participants in the TEI, who will assist in the teaching of workshops and tutorials, or in the provision of TEI-related consulting to affiliated projects and others, as described below in sections , , and (these are also listed in the budget as consultants) individuals, usually participants in TEI work groups, who will perform liaison between the TEI and other projects, as described below in section

The consultants who will assist with documentation, workshops, consulting, and liaison work are expected to be drawn in large part from a pool of qualified individuals already familiar to a greater or lesser extent with the work of the TEI. Many will have served on TEI work groups or working committees; others will have been associated with projects using TEI markup in their work. Such individuals have always been central to the work of the TEI, and there is little doubt that we will be able to find qualified people willing to do the work in question. The provision of consulting funds will make it feasible for them to do more intensive work in the preparation of extra documentation, in the teaching of workshops, and in the provision of consulting to projects, than has been possible to date.

The actual selection of consultants to perform work under this grant will be made by the editors, subject to the approval of the Steering Committee.

The office staff described represents a modest expansion of the current staff in the Chicago office of the TEI; the expansion reflects the increased need, under this proposal, for assistance in scheduling workshops, arranging organizational details with local hosts and with outside consultants, and distributing copies of the Guidelines and other documentation. Plan of Work

We propose, using the funds requested, to help ensure that the TEI Guidelines become directly accessible, and thus directly useful, to scholars involved in humanistic textual research. Our plan of work has four main areas of activity: (1) to develop appropriate written documentation, (2) to develop and teach live workshops and tutorials, (3) to create and publish a book of case studies showing the application of TEI markup to real research problems, and (4) to provide (within reasonable limits) assistance for projects in need of advice on markup and text management issues. Specifically, we intend to: disseminate the TEI Guidelines and accompanying documentation, in both electronic and paper form prepare general and discipline-specific introductory manuals on TEI markup develop materials for tutorials and workshops train workshop leaders organize and hold TEI workshops and tutorials organize a TEI Users Conference to allow users of the Guidelines to share their experiences with others publish a collection of case studies in TEI markup (the case book) provide low-level consulting for projects in need of general assistance provide more intensive consulting for selected projects using or considering the use of TEI markup prepare plans for development of model applications of SGML and the TEI markup scheme to humanities research areas take up and maintain contact with organizations and institutions relevant to the creation and distribution of electronic text

This plan does not include other concurrent activity which is expected to be separately funded: continuing technical work on the Guidelines (maintenance, preparation of revisions and extensions) development and implementation of procedures for systematic internal and external evaluation of the Guidelines

The specific activities covered by this funding proposal are described in more detail below. Disseminating Documentation

The public drafts of the TEI have been distributed free of charge both electronically and in paper form, in an effort to elicit comment from the widest possible circle of potential users. We hope to continue to make the Guidelines themselves, and the auxiliary documentation to be prepared under the auspices of this grant, as widely available as possible. Since it is not feasible to continue an indefinite subsidy for distribution of paper copies of the Guidelines, however, it will be necessary to begin charging for paper distribution in amounts sufficient to recover at least the cost of printing, binding, and mailing.

The TEI plans to distribute copies of the Guidelines and auxiliary documentation on the following basis: electronic copies will be available without charge both in TEI-tagged form and in screen-readable form (in so-called markup-free versions), by Listserv and by anonymous ftp; those without access to Bitnet or the Internet may order copies on diskette, for which a modest charge will be made paper copies will be available at a price sufficient to recover the costs of printing, binding, handling, and mailing electronic documents for use with specialized document browsing software (see below) will be distributed at prices sufficient to recover the costs of preparing and distributing them

Staff at the TEI editorial offices in Chicago and Oxford will handle orders for documentation in both paper and electronic forms.

TEI-tagged and screen-readable electronic copies of documentation will be placed on a Listserv file server at the University of Illinois at Chicago and FTP servers in England and Japan. (FTP service from Chicago will be instituted when practicable.) This ensures that any user with access to the Internet can obtain copies of all public TEI documentation without charge. The file server in Chicago is shown on the budget as a portion of the cost-sharing contribution of the University of Illinois. The servers in England and Japan are donated by their operators; they do not appear in the budget.

Some paper documentation may be published commercially, if appropriate agreements can be reached with the publishers; this may require restricting the electronic availability of the materials in question. In deciding whether to publish material commercially, the TEI will weigh the benefits of commercial distribution against the disadvantages of possibly restricted network access, and attempt to ensure the widest possible availability of the TEI's results. We will continue to investigate this matter between now and the beginning of the grant period. Preparation of Introductory Manuals

The most pressing need for many potential users of the TEI encoding scheme is a readable account of markup problems and the solutions offered by the TEI to those problems; such an account should not assume extensive prior knowledge of computers beyond basic word processing. Experience has shown, moreover, that introductory material is much more enthusiastically read if it contains examples from the reader's field.

Accordingly, the TEI plans to create introductory manuals for scholars in a variety of fields, each of which will: introduce basic problems of text representation and text markup use examples from the specific field or discipline introduce the most important elements from the TEI's core tag sets describe at least briefly any further elements particularly salient to the target discipline provide enough general understanding of SGML and the TEI to enable the reader to work independently with the text of the Guidelines in learning more about portions of the Guidelines relevant to the reader's work

As a first step toward the preparation of these discipline-specific introductions to the TEI, a generic introduction has been drafted (document TEI U1, a copy of which accompanies this proposal). Over the coming months, this document will be revised and refined on the basis of public comments and of experience using it to teach new users about the TEI. During the grant period, variants of this introductory manual will be created for at least the following distinct target groups: literary scholars textual editors dictionary makers linguists corpus-builders historians Versions aimed at other disciplines, or at more specific subgroups of any discipline (classicists, specialists in specific languages, etc.) may also be prepared if there is demand. Variations in demand may also result in the revision of the list above.

In addition to introductory manuals prepared centrally by the TEI under the terms of this grant, it is to be expected that some enthusiastic practitioners will create specialized introductions for their field, with or without prompting from the TEI. (An introduction to the TEI recommendations for the encoding of terminological data, for example, has already been prepared by the TEI work group on terminology, which has been widely distributed within the terminological community.) The TEI will assist such independent efforts as far as possible, since we do not expect to claim any exclusive right to create introductory manuals for TEI tagging.

The work of creating these discipline-specific introductions to TEI tagging will be performed in part by the editors of the TEI and in part by consultants retained for the purpose. The consultants will be expected to have adequate background in SGML and the TEI recommendations, as well as an advanced degree or experience in the discipline in question.

A separate tutorial (or set of discipline-specific tutorials) will introduce the basics of SGML, the Standard Generalized Markup Language, which provides the syntactic basis for the TEI encoding scheme. (Two basic introductions to SGML, one generic and one keyed to the interests of historians, are included among the appendices as documents TEI ED W39 and TEI ED W25.)

Given the increasing interest in automated typesetting from electronic texts, it may also be desirable to prepare a guide to the use of TEI markup for typesetting purposes, roughly along the lines of the guides to the preparation of electronic manuscripts issued by the Association of American Publishers and the University of Chicago Press. Preparation of Ancillary Materials

Other ancillary materials may be useful in making the Guidelines useful for scholars; the following are likely candidates for completion during the term of this grant: a guide to What TEI Markup Means for Your Text-Capture Project a single diskette containing an electronic form of the TEI Guidelines, sample TEI-encoded texts, and other relevant documentation a diskette (or set of diskettes) containing public-domain and shareware software for use with SGML- or TEI-encoded texts, and possibly public-domain materials to make it easier to use the TEI tag set with commercial SGML products (e.g. compiled versions of the TEI DTDs suitable for use with Mark-It, Author/Editor, and other SGML software) an electronic version of the Guidelines for use with the DynaText browser created by Electronic Book Technologies in Providence, Rhode Island possibly other electronic versions of the Guidelines or other material, for use with other browsing systems

The guide to using the TEI in text-capture projects is an attempt to address the issues raised by numerous inquiries to the TEI editors and steering committee, as is the diskette of freely available SGML-aware software.

The electronic versions of the Guidelines, by contrast, are intended in part to make the Guidelines easier to use, and in part to help demonstrate the utility of SGML encoding for the distribution of texts in electronic form. Good browsing software (such as DynaText) adapts the display of the electronic text to the computer system at hand, allows for more elaborate and complex searches than are possible with paper documentation, provides mechanisms for automatic traversal of in-text cross references and other hypertext links, and makes it possible to improve the examples, e.g. by showing both a graphic image of an original source text and one or more TEI-tagged versions of the same text, or by allowing the user to process the examples using SGML software on the local system.

In planning to distribute the Guidelines both in TEI-tagged form and in DynaText form, we distinguish firmly between the archival SGML-tagged form of the Guidelines and other forms in which they may be distributed for use with specific software. In making this explicit distinction we adopt an idea pioneered by the Perseus Project of Harvard University, which prepared its textual materials in SGML-tagged form and distributes them in a HyperCard format derived automatically from the SGML form; Perseus thus ensures that its materials will remain accessible and useful even after HyperCard ceases to be useful, and allows for the possibility of later releases using software not yet developed today. Software-specific distribution forms allow the Guidelines (or other materials) to be distributed together with software suitable for browsing, reading, and searching the electronic text. This makes the material significantly more accessible to users who may not otherwise have suitable SGML processing software handy. Since browsers can often be distributed at low cost or at no charge, the only major costs incurred are for the labor and software needed for the preparation of the browsable text.

In the context of humanistic scholarship, however, software has an exceptionally short lifespan: the oldest software still running today is unlikely to be much more than twenty-five years old, and the vast majority of programs running on computers today is less than two or three years old. It is therefore absolutely essential that electronic texts be prepared and archived in a system-independent format, rather than in a form geared to one specific piece of software. That way, the texts can always be delivered using whatever software is most suitable at the time.

We have chosen DynaText for the initial software-specific form of the Guidelines largely because it occupies a commanding position among browsing systems for SGML-encoded text. (It is used, for example, in the IBM PC versions of the Patrologia Latina and the English Poetry database published by Chadwyck-Healey.) DynaText browsers are available without charge, and run on IBM PCs under Windows, on Unix systems, and on Macintoshes. DynaText works directly on SGML-encoded text, so it will be relatively easy to import the text of the Guidelines into the system: the labor will come in customizing DynaText in ways that make it easy to use for consulting the TEI Guidelines.

Other systems for distribution of the electronic text of the Guidelines will also be considered, if the quality of the software is high enough to justify the effort and time involved. At present, we can give no exhaustive list of the browsing systems which should be considered as potential delivery systems for the Guidelines: the field is changing so quickly that any such list will be out of date by the time the grant period begins. A list will be prepared at the outset of the grant period, and candidate distribution systems will be judged by the following criteria: availability of free or low-cost run-time browsers for the system quality of text display for reading quality of indexing, search, and navigation facilities ability to handle graphic images, support dynamic processing of examples, etc. cost of software needed to create browsable documents availability of browsers on a variety of hardware and software platforms N.B. A TEI-tagged version of the Guidelines will be available in any case; the electronic distribution formats mentioned here are specially pre-processed versions for use with text browsing and text searching software. Preparation of Materials for Tutorials and Workshops

During the course of its development, various participants in the TEI have given talks, workshops, and tutorials on electronic text markup and the TEI Guidelines, ranging from forty-minute talks at conferences to week-long intensive workshops for projects involved in the creation of electronic text resources. In the summer of 1991, the TEI itself sponsored two workshops (one in Providence, Rhode Island, and one in Oxford, England, with the kind cooperation of Brown and Oxford Universities, respectively), which received a large and enthusiastic response. In the summers of 1992 and 1993, an introduction to the TEI was included in the summer seminar on electronic text sponsored by the Princeton-Rutgers Center for Electronic Text in the Humanities (CETH); the experience of the TEI workshops and the CETH seminars indicates clearly that a useful overview of the TEI can be given in one or two days, but also that a full introduction to the practice of text encoding, with practical hands-on examples, will require about a week.

As the Guidelines have neared completion, an increasing number of organizations have expressed an interest in hosting a conference session, pre-conference workshop, or other introduction to the TEI. The Steering Committee and editors cannot meet the demand by themselves, and so it is essential to create a larger group of people who feel themselves prepared to give introductory tutorials on the TEI.

As a first step in this process, the editors will prepare, with the assistance of one or more consultants, a set of materials for use in leading TEI workshops and tutorials, with special attention to the needs of: conference sessions devoted to the TEI half-day workshops full-day workshops two-day workshops one-week workshops workshop sessions on specialized topics (e.g. the application of the Guidelines to a specific discipline area, or an overview of a specific specialized portion of the Guidelines), which may be used in the course of workshops for specialized groups

In addition, the editors will prepare, with help from one or more consultants, a , which will function as a sort of instructor's manual for the use of the workshop materials. Training Workshop Leaders

After a first version of the workshop materials is produced, we will hold a workshop for training workshop leaders and TEI consultants. In five days, we will review the workshop materials, discuss various possible ways of structuring a TEI tutorial or workshop, and demonstrate various methods (group tagging, individual work, software demonstrations) of introducing potential users to TEI tagging.

This workshop will draw on the extensive experience of the editors, steering committee, and others in teaching TEI tutorials over the past few years.

It is expected that this workshop will take place sometime in the first six months of the grant period. Following this workshop, the materials prepared for it may be revised in the light of comments and suggestions made by the participants --- who will, after all, be using the materials in their own future workshop activities.

Participants in this workshop for workshop leaders will be expected to assist with the organization and teaching of other TEI workshops, as described in the next section, or with the provision of consulting services to affiliated projects, as described below in section . TEI Workshops and Tutorials

Using the materials prepared, the steering committee, editors, and the trained workshop leaders will teach a series of TEI sessions and workshops, both independently and in conjunction with other academic conferences. The organizations represented on the TEI Advisory Board will be asked to schedule sessions or pre-conference workshops for their membership, and to handle local organization and financing of the workshops; other organizations expressing interest will be accommodated as far as possible. Since some organizations will be unable to cover travel and consulting costs for the teachers of the workshops, funds are requested under this proposal to cover those expenses for at least some conference-associated workshops.

In recent years, sessions on the TEI have been scheduled at the conferences of its sponsoring organizations, as well as at the Society for Biblical Literature, the annual conference of the International Computer Archive of Modern English, and the Modern Language Association of America. One- or two-day workshops have also been sponsored by the Network of European Corpora (a collaborative effort to develop a coordinated set of standard reference corpora for the modern European languages) and (as already noted) the Center for Electronic Texts in the Humanities. Future sessions have been scheduled for the next annual meetings of the American Society for Information Science, the Social Science Historical Association, and other organizations.

While it is difficult to predict exactly how many organizations will sponsore a TEI tutorial or workshop in connection with their conference, or how many of them will require TEI subsidies for the travel costs of the workshop leaders or session speakers, it seems likely that at least the majority of the organizations on the Advisory Board will do so, as well as some others. Accordingly, the budget includes a request for travel funds and consulting time to cover fifteen such conference presentations over the two years of the grant.

The TEI will also sponsor two week-long workshops itself, tentatively scheduled for June, 1995 and 1996. Participants in these workshops will be expected to cover their own travel and subsistence costs, and may be charged a small fee to cover local organizing costs. If funds allow and demand is high, the TEI may sponsor more than two such workshops. TEI Users Conference

During the second year of the project (approx autumn 1995), we will organize a small conference in Chicago or elsewhere in North America to allow users of the TEI to report on their experiences, and allow the TEI to report on the progress of its continuing revision and development work. We will pay for the attendance of TEI participants, and will set aside a small fund to subsidize attendance by people who would otherwise be unable to attend, but most participants will be expected to pay their own way. TEI Case Book

Selected papers from the TEI User Conference, supplemented if necessary by other papers, will be collected into a book of case studies in the application of the TEI to problems of text encoding and scholarly research. Some reports from our consulting with affiliated projects may also be included in this collection. Low-Level Consulting

We will continue to provide, as we have done during the development period, basic information and advice to researchers concerned about how to put the TEI encoding scheme to use in their own work, and trying to work out the implications of a scheme like that of the TEI for their data capture projects. With the wider dissemination of the Guidelines, this level of consulting is expected to take an increasing amount of time, but as knowledge of the TEI and its implications becomes more widely spread, we expect some of the simpler questions to be answered by local experts before the inquiries reach the TEI, so that by the end of the grant period, the load of such queries may be expected to decline again. Intensive Consulting with Affiliated Projects

One of the most successful parts of the TEI's work during the development effort involved close collaboration with a small group of affiliated projects, who agreed to test the draft Guidelines on their data and to report on the results, in exchange for receiving some training and consultation on problems of text markup. The affiliated projects thus provided an essential link to practical issues, and an essential check on the practicability of the Guidelines, which benefitted the TEI a great deal. The affiliated projects report that they, in turn, benefitted from the intensive discussion of text representation issues during the week-long workshops for affiliated during the summer of 1991.

We propose to continue the practice of consultation with affiliated projects during the dissemination phase of the TEI, and to make it an integral part of our attempts to spread knowledge of the encoding scheme into the practice of humanities research, as well as of our efforts to test the TEI markup language against real research problems. Since it requires real resources to provide the kind of consultation, training, and site visits which made our work with affiliated projects useful in the past, the number of affiliated projects must be limited, and resources must be devoted to those projects which offer the highest likelihood of offering interesting and fundamental problems for text markup posing problems common to many projects, so that the experience of the affiliated project will be enlightening to others being able to devote sufficient resources to the cooperative effort to make it yield useful samples or experience bringing useful background or skills to the effort

At the outset of the grant period, or a little before, we will post a public request for proposals, inviting interested projects and individuals to contact us with descriptions of their work, the nature of their interest in the TEI, and the specific work they hope to accomplish with the TEI as an affiliated projects. From the proposals received, the Steering Committee and editors will select eight to fifteen projects as affiliated projects. Each affiliated project will be allowed to send one representative, to be funded by the TEI, to one of the TEI-organized workshops, and will be assigned a consultant from the pool of trained consultants and workshop leaders, to assist them in their TEI-related work. We budget, as well, for a site visit of two days, for on-site consultations; further site visits should be funded by the affiliated projects themselves, although in exceptional circumstances the TEI may subsidize them, funds permitting.

Each affiliated project and consultant will be expected to produce a short report on the work undertaken, which will be published as a technical report and may also be included in the TEI case book described above. Model Applications

In the long term, the success of the TEI in helping humanists with their research and teaching depends on the existence of visible models of ways of using TEI encoding in practice. Much of the work with affiliated projects is designed to produce such models, but in some instances, further efforts may be appropriate. Most important, the level of consulting and cooperative work feasible with the affiliated projects makes it possible to discuss the best ways of using existing software to achieve the goals of a project, but does not suffice for efforts at developing new software or new combinations of software for scholarly purposes.

The paucity of suitable, readily available, affordable software, however, is one of the major problems facing any humanities research project, and one of the key barriers to wider adoption of SGML in the humanities research community. It is essential that we take steps to address this problem, and we propose to do so in several ways.

First, the TEI will continue to encourage commercial vendors of SGML software and services to support the kinds of processing most needed for humanities researchers. Almost always, the features required by humanists have commercial application as well, but not all vendors perceive the commercial viability of, say, text-critical apparatus without it being pointed out to them. Version control is a major problem for technical publishers and organizations like the European Community, which publish documents in parallel versions in multiple languages. The same processing needed to support the encoding of manuscript variants or parallel texts may be used to support multiple language versions of the same document, or version control for technical documentation.

Second, the TEI will cooperate in ventures such as the Text Software Initiative (TSI), an international venture to encourage the production of freely available software for processing SGML texts. The TSI was begun by one of the founders of the TEI, and several members of the TEI Steering Committee sit on the TSI steering committee as well.

Finally, we will seek for opportunities to collaborate actively with other interested projects or organization in the development of specific pieces of software usable in specific types of humanities projects. Such model applications will demonstrate, in a concrete environment, the use of TEI markup in real research projects, and will adapt or develop the software needed to make TEI markup work successfully in those environments.

The development of model application software will in general be separately fundable, and so no funds are included in the budget of this proposal for any model application development. However, the negotiation of agreements with other projects, and the preparation of concrete plans for such work, are expected to require serious effort, and therefore we propose, under this grant, to engage in planning activities for model application development, including: selection of suitable areas for efforts of this kind discussions with projects in those areas, seeking willing partners for a model application development effort preparation of preliminary descriptions of the model application, and of plans for its development under appropriate circumstances, performance of simple feasibility studies for the model application being considered It is expected that most projects interested in cooperating with the TEI in developing a model application will also be interested in working as affiliated projects, and so much of the work in studying the feasibility of model applications will take place within the framework of the affiliated-projects program, and is not separately itemized within the budget. Liaison with Other Organizations

As the TEI has approached the conclusion of its development efforts, outside interest on the part of SGML vendors, similar projects by other groups, and standards organizations has grown. It is clear that the TEI can have perceptible impact on the development of a wide variety of standards in the coming years, if we can sustain the effort of informing interested parties about the TEI before they have gone to the effort of inventing (yet again) another ad hoc system for encoding text. We propose, therefore, to use funds from this grant to ensure adequate exchange of information between the TEI and: ISO working groups on SGML, document processing (DSSSL), hypertext, and character sets the European Expert Advisory Groups on Linguistic Engineering (EAGLES), which are developing guidelines for the encoding of texts for projects funded by the European Community (taking, in some cases, the TEI recommendations as their starting point) Internet-based efforts to define standards for extended mail formats (MIME) World-Wide Web, a network-based hypertext effort the International Committee on Accessible Document Design (ICADD), an international effort to devise SGML tag sets suitable for the delivery of text to sight-impaired readers Recording for the Blind, a major North American provider of materials for the print-impaired reader, which is continuing to develop its internal standards for the creation and distribution of electronic text efforts to create large collections of data for use in linguistic research and engineering (including inter alia the Consortium for Lexical Research and the Linguistic Data Consortium in the U.S., the European Corpus Initiative and the Network of European Corpora in Europe, the international Data Collection Initiative sponsored by the Association for Computational Linguistics) efforts to provide access to electronic text for research, especially over the network (including the Center for Electronic Text in the Humanities, the Coalition for Networked Information, the Oxford Text Archive, and other data archives) For the most part, expenses under this heading will go to fund travel by TEI representatives to attend meetings of the other organization; other expenses will involve the distribution of relevant TEI documents to people doing the technical work and the acquisition of technical documents from the other body and their redistribution to the appropriate participants in the TEI. Individuals and their institutions will be expected to donate their time for technical liaison work; no consulting fees will be paid. Concrete Results: Final Products and Dissemination

The specific documents to be produced and activities to be funded by this grant are described in more detail in the preceding sections; in summary, the documents to be produced are: revision of the generic introductory manual on TEI tagging revision of the TEI introduction to SGML six or more specialized discipline-specific introductions to TEI tagging in several disciplines a guide to the use of TEI markup for electronic manuscripts a document describing the implications of TEI markup for text capture projects of the types commonly funded by NEH a diskette containing the TEI Guidelines and other documentation in TEI-tagged form a diskette with public-domain and shareware software for SGML processing a DynaText version of the TEI Guidelines list of browsing systems to be evaluated as delivery vehicles for distribution of the TEI Guidelines an evaluation of selected browsing systems as potential delivery vehicles for electronic versions of the Guidelines complete sets of materials for teaching TEI workshops and tutorials ranging in length from three hours to a week a Guide to Teaching TEI Tutorials and Workshops a book of sample uses of TEI markup to encode specific types of text or to solve specific research problems reports on results of TEI's consultation with specific projects reports on TEI's liaison with other institutions and projects

The activities to be performed are: provide infrastructure for the distribution of paper and electronic copies of the Guidelines and other documentation (office staff, electronic file servers) select individuals willing and able to serve as workshop leaders and project consultants schedule and hold a workshop for leaders of TEI tutorials and workshops and consultants working with affiliated projects schedule and hold a series of conference sessions and/or TEI workshops in conjunction with meetings of major professional societies schedule and hold two to four TEI-sponsored workshops of two to five days' duration, held independently of other conferences or meetings schedule and hold a TEI Users Conference for the exchange of user experiences provide information and general advice to researchers who inquire about the use of TEI markup in their project select projects interested in intensive work with the TEI as affiliated projects exploring problems in applying the Guidelines to specific areas provide more intensive consultation with these affiliated projects maintain technical liaison with relevant organizations and projects Timetable

Much of the activity covered by this grant proposal will take place more or less continuously throughout the grant period: handling distribution of the Guidelines and other material scheduling and holding conference-related workshops and tutorials on the TEI providing basic consulting and advice for scholars interested in using or learning about SGML and the TEI Guidelines consulting with affiliated projects and reporting on results maintaining liaison with other relevant organizations and projects and reporting on results These activities will not be listed separately under each time period. July-December 1994 Revise generic introductions to SGML and TEI tagging. Issue call for affiliated projects and for TEI workshop leaders; select affiliated projects and TEI workshop leaders and consultants. Produce the following materials: one discipline-specific introduction to TEI tagging guide to the implications of TEI markup for text-capture projects diskette with electronic TEI-tagged versions of the Guidelines and other documents list of browsing systems to evaluate as candidate delivery vehicles first version of teaching materials for TEI workshops and tutorials Hold the workshop for leaders of TEI workshops and tutorials; begin revision of workshop teaching materials. January-June 1995 Produce the following materials: one discipline-specific introduction to TEI tagging guide to using TEI tagging for electronic manuscripts disk with public-domain SGML and TEI-aware software revised version of TEI workshop-teaching materials Begin work on DynaText version of the Guidelines. Organize and hold week-long TEI workshop (June 1995?). Begin organization of TEI Users' Conference. July-December 1995 Produce the following materials: DynaText version of TEI Guidelines evaluation of electronic browsing systems two discipline-specific introductions to TEI tagging Hold TEI Users' Conference. Begin work on TEI case book. January-June 1996 Produce the following materials: two discipline-specific introductions to TEI tagging TEI Case Book: Examples of TEI Tagging in Use Hold week-long TEI workshop. Notes on the Use of Automation Technology Rationale

Computing machinery will be used in the conduct of the work covered by this grant because: electronic mail speeds correspondence and the exchange of drafts among participants, word processing systems make the production and review of documents by multiple authors faster and more reliable than exchange of paper drafts, electronic conferencing provides one of the best methods of publicizing this initiative among the interested community, and electronic browsable versions of SGML text offer one of the best methods of demonstrating the power of SGML to potential users.

Computers are, in addition, part of the normal working conditions for the participants in the project and are used in their daily work. The computational load on the consultants is unpredictable, and costs are expected to be borne by their host institutions. Funds are included for the editorial sites in Chicago and Oxford, which will handle most of the documents involved in the project; these funds will be contributed by the University of Illinois at Chicago (UIC) and the Oxford University Computing Services (OUCS). Hardware

No hardware is devoted exclusively to work on this project. The editors and members of all committees use whatever resources are normally available to them. These range from IBM mainframes running VM/CMS through Vax systems running VMS, desktop workstations of various persuasions mostly running Unix, to IBM PCs and compatibles and Apple Macintoshes. Software

Text processing software used for the project is similarly various. Documents are prepared by the editors in SGML-tagged form, from which various derivative forms are prepared by automatic processing (including forms suitable for processing by Waterloo Script/GML and LaTeX, which the editors use for formatting and printing).

Although it is anticipated that software may be developed to process data encoded in the format developed by this project, such software development will take place in other projects which will have their own funding requirements. No formal software development is anticipated as part of the central activities under this grant. Costs

Costs for computer use included in the main budget cover costs at the central editorial site at UIC and costs of the associate editor's work in Oxford; these estimates are based on past billings on the machinery in question for work on this project. It may be noted that the machine costs given compare favorably with those in many comparable centers at institutions around the country. Financial Information Funding Funding History

The TEI began with a proposal for a five-phase project: planning and high-level design low-level design and drafting circulation and review of drafts formal approval by an advisory board of participating organizations publication and maintenance The first phase of the project was funded by NEH under grant RT-20880-87 to Vassar College, with support also from Vassar. The next three phases of the original project plan have been funded in part by NEH, under grants RT-21016-88 and RT-21193-1990; they have also been funded in part by a grant from the Andrew W. Mellon Foundation and a series of contracts issued to the University of Pisa by the Commission of the European Communities, under which the contractor coordinates the participation of European researchers in the work of the Initiative. Support for related activities has come from the Social Science and Humanities Research Council of Canada.

The work done to date has also been supported by donations in kind and direct or indirect subventions from the University of Illinois at Chicago, the Oxford University Computing Service, the University of Arizona, the University of Oslo, Queen's University (Kingston, Ont.), Bellcore (Bell Communications Research), the Istituto di Linguistica Computazionale (C.N.R.) (Pisa), Vassar College, the Princeton-Rutgers Center for Electronic Texts in the Humanities, EDR (the Japan Electronic Dictionary Research Institute, Tokyo), and Ohio State University, as well as the employers and host institutions of the members of the TEI working committees and work groups. A variety of software firms have assisted with software acquisition. Funding Plans and Prospects

This proposal requests partial funding for the dissemination of the work done to date, principally for dissemination within the U.S. Funds for European dissemination in this work will be sought from European sources; our current contract with the Commission of the European Communities provides specific funding for dissemination activities, and a renewal will be requested at the appropriate time.

Support is now (at the time of writing) being sought from other agencies for work other than that covered in this proposal (notably continuing technical work on the Guidelines and their extension to areas not now covered, and formal evaluation of the Guidelines). At present, a pre-proposal description has been circulated to relevant parties within ARPA (the Advanced Research Projects Administration) and NSF (the National Science Foundation), both of whom have some interest in electronic representation and manipulation of text. We will keep the Endowment informed of developments in this area. Budget

This section follows the budget forms item by item, indicating the rationale for each expenditure and the method used to estimate each cost. The budget covers only the activities directly supported by this grant, and excludes other funds.

The work of the project will be supported for the most part through the University of Illinois at Chicago (UIC), and in part through a subcontract with Oxford University Computing Services (OUCS). Salaries and Wages

We request funds to cover 50% time for C. M. Sperberg-McQueen and Lou Burnard, a full-time office manager / program coordinator, and one student assistant to help with organizational matters. Lou Burnard is to be paid through a subcontract with Oxford University Computing Services (see letter of commitment, attached); the other staff members will all be paid through UIC.

The editors' responsibilities under this grant will be to draft the documentation described above in section , or to commission outside consultants to do so; to supervise the distribution of the Guidelines and other materials; to prepare or supervise the preparation of the DynaText and other electronic versions of the Guidelines; to prepare the package of teaching materials for TEI workshops and organize and run the TEI workshop for consultants and workshop leaders, all in conjunction with outside consultants; to supervise the scheduling and organization of TEI workshops and tutorials and the TEI Users' Conference; to edit the case book of TEI examples; to select, in conjunction with the Steering Committee, affiliated projects for intensive project-related consulting; to select, with the concurrence of the Steering Committee, the consultants who will assist in teaching workshops and in consulting; and to supervise (and participate in) the teaching of TEI workshops and the consulting with affiliated projects.

The office manager / program coordinator is responsible for managing the orderly distribution of the Guidelines and other materials, assisting with the proper formatting and printing of paper documents, maintaining project-internal databases, routine correspondence, handling routine organizational tasks associated with the scheduling of workshops and conferences, ensuring that payments for travel costs and consulting time are properly handled within the university accounting system, overseeing the TEI office in Chicago, and supervising the student assistant.

The student assistant is responsible for assisting with database maintenance, mailing, bookkeeping, word processing, and general office tasks.

Salaries: 1994-95 1995-96 C. M. Sperberg-McQueen (1/2 time) $22 785 $23 924 Lou Burnard (1/2 time) 21 500 21 500 Office Manager / program coordinator 28 000 29 400 Student assistant 9 600 10 080