Symposium introduction

Evan Owens, the symposium chair, will introduce the context of up-translations and up-transformations, identify critical issues in workflow solutions, and summarize the themes and topics to be covered by the conference speakers and in the demos. Content enhancement by up-transformation and/or up-translation has historically been an important task and component of markup language technology. Requirements keep evolving; this has been and will remain an important and perpetual topic.

Up and Sideways

A conversion of hundreds of Rich Text Format documents to highly structured XML is always going to be a challenge and a showcase of XML technologies, even if you are excluded from a number of them. This paper is a case study of one such conversion, dealing with migrating huge volumes of legal commentary, more specifically the classic standard text Halsbury's Laws of England, from RTF to XML so new editions can be authored and published in XML to various paper and online publication targets. While describing the migration approach in any detail would probably require a book-length paper, this attempts to highlight some of the challenges and their solutions.

Uphill to XML with XSLT, XProc … and HTML

HTML is a widely familiar vernacular for ad-hoc representation of documents, and can be useful as a staging ground for decomposing and breaking down the more complex operations in uphill data transformation. HTML, syntactically well-formed and maintained within XML pipelines with well-defined interfaces, can usefully join XSLT and XProc to provide for a complete up-conversion or data-enhancement pipeline – especially when the ultimate target is semantically richer than HTML. In a project based on this approach, lessons learned include: “Many steps may be easier than one”; “If it doesn't work, try it the other way around”; and “Validation is in the eye of the beholder”.

Looking for Rumpelstiltskin

Up translations are a black art; the stuff of conjurers and tricksters. The focus of this case study is the migration of unstructured and semi-structured formats to XML for a healthcare information provider with more than 20 different product offerings targeted to healthcare facilities, insurers, and practitioners. This paper examines the approaches taken to spin their unstructured and semi-structured content into XML and the challenges encountered along the way.

Rebuilding a Digital Frankenstein by 2018

Digital editions were young in the 1990s, and the expansive possibilities of hypertext in that decade sharply distinguish early digital editions from the productions of our moment. The accessibility and simplicity of early HTML code made for innovative experiments with the size of a “page” and the way one might handle displays of variants, before “diffing” tools like the Versioning Machine and Juxta came to define how we usually imagine the digital comparison of texts. This paper investigates the serious problems and vexing potentialities of “up-translation” when standards change, concentrating on work underway on a Bicentennial Frankenstein project. Our project is to produce a new, freshly collated digital edition in TEI based on the Frankenstein texts digitized by Romantic Circles, and incorporating a little-known publication of 1823 together with 1818 and 1831 versions currently represented. Readers in the past century are likely to have encountered either the 1818 or the 1831 edition but not the 1823, and we think that folding this text into our collation may help us to understand more about when and about how gradually some of the major alterations in the 1831 text (for example, to Victor Frankenstein's family members and the compression and reduction of a chapter in part I) occurred. The three print editions will be compared in parallel, and we will incorporate pointers to the Shelley-Godwin Archive’s edition of the MS Notebooks. To prepare the collation we returned to the simplest original form of the current edition, the Pennsylvania Electronic Edition prepared by Stuart Curran and Jack Lynch in the 1990s. Exploring that edition exposes the ambitious intellectual scope of early web editions and raises important questions about how we built editions then vs. now. We do not build a new edition to replace the earlier work, which yet “lives” and is available on the web. But our work is a fresh start, not a seamless integration. This particular project’s encounter with an impressive early hypertext edition raises more general questions worthy of reflection towards theorizing the up-transformation process: How do we understand the relationships among generations of digital editions? What aspects of the old hypertext editions (or editions in formats not consistent with our own) transcend or exceed the structures we currently consider sustainable? What perspective might a thorough review of the first still extant hypertext editions contribute to our scholarly editing practice now?

Finding the Tipping Point in Automated Markup During Up-Translation

Up-translation can be accomplished using both automation and manual techniques. For most complex content, complete automation can introduce errors or miss content that requires tagging, but manual tagging is time consuming and potentially error prone. The best results can be obtained by finding a middle ground between automation and manual tagging. However, finding that middle ground is, itself, a challenge, and addressing that challenge requires a careful balancing act of investing in software development for automation, automatic flagging of suspect cases for manual review, and designing a tagging and quality-assurance workflow that is both robust and efficient. This paper discusses the inevitable inconsistencies, ambiguities, and “gotcha” moments that are encountered when up-translating scholarly manuscripts to models such as JATS and BITS, and provides recommendations for balancing automation with manual review.