Diple, modular methodology and tools for heterogeneous TEI corpora

Digital humanities, TEI, Relax NG, XSL, PHP, CeCILL-C (LGPL)

Documentation
Forge (SourceSup)
Sources, svn checkout https://subversion.cru.fr/diple/trunk
Mailing list: diple-dev@cru.fr

What is Diple?

Diple is a collection of modular tools, and a methodology, for the publication of XML corpora, which allows a streamlined editorial workflow for each new project.
The logo combines a kleene star (*) and a “diple” (>), a sign from the classical editorial tradition.

Examples

Workflow: Publishing XML-TEI Corpora

We developed a shared TEI schema for all ENC publications (diple/schema/).
An XSL engine (diple/transform/) handles all corpora conforming to this shared schema.
An editor can focus on a new project's specificities: the corpus' XML file, and short corpus-specific RNG and XSL files (corpus/).
Diple's PHP and JS libraries (diple/php/ and diple/js/) generate the final site.
Diple's XREM module (diple/xrem/) generates schema documentation.

Module XREM: Documenting XML Files (Relax-NG, XSL, XML samples)
`diple/xrem/`

A precise schema requires the scientific editor to document and make explicit editorial choices...
... and requires encoders to conform strictly to the edition's rules (a real use of XML-Validation).
XREM takes the schema documentation (a set of Relax-NG files) and produces an HTML documentation cross-referenced with the TEI docs.
XREM can also document XSL stylesheets.
XREM is a set of XSL and Javascript files that can be run on a personal workstation.
By releasing the docs and XML files for an electronic edition, a publisher goes a step further in diffusing important scientific information.

Selected Features

Reader-friendly presentation for editors (and not only programmers).
For Relax-NG : distinguishes elements by their attribute values (<div type="tradition">, <div type="transcription">).
for XSL : indexes input and output elements to quickly find a template.
Documents any XML using XML comments  (no need for a proprietary namespace).
XSLT 1, permits transformations in web browser.

Diple XSL-Engine
`diple/transform/`

XSL-Engine based on our schemas supports those TEI elements specific to critical editions.
Generates the visual features of a critical edition: multiple levels of footnotes (historical and philological), critical apparatus (presentation of variants linked to the manuscript tradition/witness list), index, table of contents ...
Shared schemas and transformations allow the editor to concentrate on corpus-specific elements.
XSLT 1, permits transformations in web browser.

Diple PHP Objects Library
`diple/php/`

Library of PHP objects that produce the front-end web site.
No dependencies between objects – this is not a framework.
Many objects are useful in other contexts (outside Diple).
Selected objects: HtmlInc (extraction of metadata and body...), Hilite (highlight regex passed as an argument), Web (manipulation of http headers...).

Integrating external research tools

Diple divides publication and search.
Each corpus has its own logic and needs its own search engine and external research tools.
Search results included inline in the electronic edition.
Relational databases (MySQL, SQLite), fulltext indexer (Solr/lucene), or concordancers (PhiloLogic) have already been plugged in.

What is a diple ?

A marginal mark of this form >, used by the ancient grammarians to indicate various readings, rejected verses, beginning of a new paragraph, etc. (Oxford English Dictionary, « diple »)

In paleography, a critical mark like a Υ or Λ laid on its side , used as a mark of a paragraph, the change from one speaker to another in a drama, different readings, rejection of a reading, etc. (Century Dictionary, « diple »)

διπλῆ, (διπλοῦς) a marg. mark used by Gramm. (<*>), to indicate vv. Il., rejected verses, etc., and, in dramatic poetry, a new speaker (Liddell Scott, « διπλῆ »)

Signum in libris præsertim Ecclesiasticis ad distinctionem appositum. Vide Isidorum Orig. lib. 1. cap. 20. (Du cange, « DIPLE »)

Gr. διπλῆ (diplè) « ligne double, à deux branches ». Signe formé de deux lignes obliques formant un angle aigu dont la pointe est le plus souvent dirigée horizontalement (IRHT, Ædilis, Lexicon, « diplè »)

Signe (<,>) servant aux critiques grecs à indiquer des variantes dans les vers d'Homère ou à noter des vers suspects, et servant aussi à indiquer les citations des saintes Écritures. (Littré, « diple »).

Credits

Authors: F. Glorieux et V. Jolivet, with contributions by G. Bardet, C. Desenclos and C. Hurel
Funding: École nationale des chartes, TGE Adonis (for odt>tei module)
Acknowledgement: Comité Réseau des Universités for providing SourceSup forge
Diple is available as open-source (CeCCIL-C / LGPL).