Main »

Tools For Collaborative Scholarly Editing Over The Web

Tools for Collaborative Scholarly Editing over the Web?

These are conference notes on a workshop put on by Peter Robinson at the University of Birmingham Institute for Textual Scholars and Electronic Editing. The conference is being Tweeted with the subject #itsee . I'm writing this conference report live so is incomplete, rough and biased.

Interedition projects

Joris van Zundert and a colleague presented first about the e-Laborate environment which is an easy-to-use online scholarly editing environment that allows one to see a manuscript page and then transcribe and annotate it. They then talked about the Intereditions project and the development of "microservices".

Dot Porter of the DHO (an Interedition partner) talked about Text Image Linking Environment and Image-Based Editing. TILE is a project funded by the NEH building on AXE from MITH. She gave a good overview of image-based linking projects and their limitations. She argued that images should be an important part of scholarly editing. It is not quite clear what TILE will be. It sounds like it will be modular and can be linked into other projects. She also sees TILE as a community whose tools will support multiple needs. Show showed a tool that automatically links images regions (words) to text.

Andrea Scotti of the Fondazione rinascimento digital talked about the PinakesText which is one of the models for a larger project called Pinakes which somehow generates interfaces and publications automatically. He walked us through an ontology and use case for annotations.

National approaches

Roger Osborne presented on AustLit (The Australian Literature Resources) and the new project Aus-e-Lit. They are building a tool, LORE, that is a Firefox plug-in that lets users open an editor where you can create relationships between objects that produce RDF graphs that are then stored for searching. The idea is, for example, to let people add annotations to AustLit works.

Steven Hayes presented on "Working with the Harpur manuscripts". Harpur was possibly the first (white) Australian poet. He showed their annotation approach where again, they want to make it easy for people to add annotations that can be stored in a database.

Yin Liu presented an overview of Canadian edition projects. She asked why most individual editions are edited by a single person. Does the idea of collaboration stop at editing a particular text? Most Canadian collaborative projects are collectives of individual projects. Are we limited by our academic habits? She suggested we look at small student projects and learn from how digital natives collaborate. She then reflected on collaborating with computer scientists and how she has learned from working with a computer scientist who challenges habits and ideas.

I then spoke about various tool projects (TAPoR, TAToo, Voyeur, and JiTR.) My theme was Reinventing Wheels and the reasons why we do reinvent things.

Approaches to universal solutions

Neel Smith talked on "Targetting infrastructure, not tasks" and the CITE architecture for identifying and retrieving objects. Infrastructure outlasts tools. We want to develop tools for infrastructure, not tasks. We want standards-based specs independent of implementations. We want to work with persistent, citable data. We want to work with specifiable machine actions. We will never get consensus on detailed ontologies. Neel feels the place we can agree is around citation - identity for citation is something we should agree about so we can cite each other's stuff. He quoted David Wheeler, "All problems in computer science can be solved by another level of indirection." The ultimate foundational functions are identification and retrieval (citation actions.)

The CITE (Collections, Indexes, Texts, Extensions) architecture developed an abstract model of canonically citable texts. They have a syntactically and semantically specified notation and an implementation called CTS (Canonical Text Services). I liked how CTS uses Google apps to provide services on the cloud. See http://cts3.sourceforge.net/

Julianne Nyan and Oliver Smidt talked about TextGrid project. They have developed a client on Eclipse that can work with a service grid and data grid. There is a working version of the client that you can download. This strikes me as a very ambitious national project to provide one grid for all.

Federico Meschino talked about ontologies from the philosophical sense to "specification of a conceptualization to description logic. There were excessive expectations for the Semantic Web and ontologies. They don't solve everything, but they can be useful. An edition is a set of facts about something that are recorded. Examples are things like FRBR and CIDOC-CRM. Fede is adapting CIDOC-CRM to textual history.

Peter Robinson followed up talking about the difference between an abstract text and actual carrier. Like the CITE project they are developing a URN model for citing things.

Project and community based approaches

Tamara Lopez with CeRch/CCH, King's College London talked about TEXTvre (pronounced "Texture") which is working with TextGrid to experiment with a virtual research environment at Kings and in the UK. TEXTvre is built around editions building. They are "need oriented" where they add middleware when scholars say they need it. They are looking at identified UK based grid services like the National Grid Service and HPC services. Some of the issues they are looking at include: public vs private work, what do scholarly editions mean, what are the roles in a project, and what editorial work still needs to be paper based.

Karsten Kynde of the Kierkegaard Research Centre talked about their online edition of Kirkegaard's work. He talked about the problem of explaining the difference between interface and data - "A book is an interface to a text."

Friday Sessions

On Friday we had more informal conversations laying out problems. Some that came up:

1. National silos. Why do we do things as national projects rather than use projects elsewhere. The answer is politics, but we can build national things that share data.
2. Ontologies came up. Do we want to get involved in massive ontology standardization projects or just share them. Is there a way to have stock ontologies which project build customizations on.
3. Collaborations should be focused.
4. How to enable social research? How to bring in the larger interested community? We talked a lot about social research. We mentioned examples like Gutenberg, an Australian newspaper project, and so on.
5. Do we know ourselves and our academic practices? I personally think one of the contributions of the digital humanities is around the confrontation of formalization and academic practices. We have been forced to think about people and their practices together.
6. More and more funding is directed to projects and not to the general system and scholarship(s) (in both senses of the word.)

We came up with some suggestions

1. Stop all funding to all projects.
2. We should fund development of tools that can be used by many.
3. Projects should not claim a tool they build for themselves will be usable by everyone.
4. Fund initiatives to allow the data made by the ools to be found and used by everyone.
5. Ontologies. We need to better describe the knowledge in projects. But ontologies may not be infrastructure.
6. All data should be shared with a creative-commons share-alike license.
7. Separation of concerns - we need to separate functions so that researchers can collaborate and technologists can. True collaboration should be across borders. It is about collaboration and collaboration is not talking, but working together. We talked a lot about collaboration and its limitations.
8. We talked about boycotting libraries that refuse to put stuff out under creative commons licenses.

I presented adapted theses from Computing with the Infrastructure at Hand:

1. It is people that matter not funding or things.
2. There is no rush. Go slow and thoughtfully and do it again.
3. We have more than we think we have.
4. We have to help each other across the usual borders.
5. We are not just in the university.
6. We can always do something that matters.
7. Know ourselves.

---

Thoughts

I'm struck by how difficult real collaboration and interoperation is. We all call for it, but don't do it except when paid to. That said, I think the time has come when some will actually try to make it work. It is worth enumerating some of the ways we do collaborate:

• The most common is through ""conferences"". We go off for a year, do a lot of work, and then present it at meetings and conferences. This can be a very efficient way of keeping projects loosely informed of each other.
• ""Collectives"" is how many projects organize participants so that everyone works alone (editing separately) communicating regularly and standardizing. Yin Liu talked about this and wondered whether it is academic habits that keep us from deeper collaboration.
• ""Web Services"" can be used to make utilities/tools available for others through a standard interface. TAPoR took this approach.