The goal of this workshop is to share ideas, common issues and principles about di models and algorithms, change tracking, collaborative editing and document versioning. We want to look at these topics from di erent perspectives and want to understand which are the most common issues and which are the peculiarities of each domain and each approach.
The core idea is that authors collaborating on textual documents should have at least the same tools that software engineers use when collaborating on source code. With some important di erences mainly due to some uncertainty and variability (for instance, the authors might not trust each other and might require a third-party validation) and lack of rigorous validating techniques (like compilers for software engineers). More important, non-technical users are not expected to have the same expertise of software engineers in using versioning control systems. The talk envisions a new This work is licensed under the Creative Commons AttributionShareAlike 3.0 Unported License (CC BY-SA 3.0). To view a copy of the license, visit http://creativecommons.org/licenses/by-sa/3.0/.
The core sessions of the workshop are on research papers. We received 12 submissions from all around the world, among which 7 papers were selected after a single-blind review process. The papers cover both practical and theoretical issues and can be clustered around three main topics: XML change management, understanding the evolution of nontextual documents and data structures and distributed collaborative authoring.
The rst session deals with the management of changes in XML documents.
In Merging Uncertain Multi-Version XML Documents, the authors focus on versioning uncertain XML documents. The problem is very challenging considering that unreliable information exists for each contribution and that some contributions cannot be trusted or merged at all. The paper presents a reliable and fast algorithm for merging versions in such a scenario, together with its proof of correctness. It is part of a larger framework, which will be presented at the main conference.
The paper An Algorithm for Transforming XPath Expressions According to Schema Evolution deals with evolving XML schemas and documents. It studies how to automatically update queries on XML documents, when these documents change to meet changes of their validating schemas. The authors present a novel algorithm based on tree automata, together with some experimental results. Though limited to only few XPath axes, the algorithm is very e cient and extensible.
The second session focuses on documents and data structures that are more complex than simple text.
In The Concept Di erence for EL-Terminologies Using Hypergraphs the authors focus on automatic detection of the logical di erence between ontologies. The logical di erence is de ned as the set of queries that produce di erent answers on the ontologies being compared. Their approach consists of modeling ontologies as hypergraphs, calculating simulations between hypergraphs and converting them back to differences between ontological axioms. The paper presents a theoretical and solid work, and anticipates possible extensions to richer logics.
The paper Staged Evolution with Quality Gates for Model Libraries deals with the evolution of model libraries. The authors put forward their quality staged model evolution theory for model libraries. Their theory is founded on evolution graphs, which o er a structure for model evolution in model libraries through evolution steps. These evolution steps eventually form a sequence, which can be partitioned into stages by quality gates. Each quality gate is de ned by a lightweight quality model and respective characteristics fostering reusability.
The paper Identifying Change Patterns in Software History focuses on di ng and versioning source code. The overall goal is to identify patterns of code changes, in order to better understand how a given codebase has evolved. The authors propose a layered approach that works on the AST (abstract syntax tree) representations of versioned les and combines a tree di algorithm and similarity grouping techniques to cluster low-level changes into higher-level patterns. The approach requires a few customizations to also work on other programming languages. Experimental analysis of two Java projects are also presented in this work.
The last session of papers is on distributed collaborative authoring.
The paper Concurrency E ects Over Variable-size Identiers in Distributed Collaborative Editing tackles the problem of building distributed editors with CRDT (Con ict-free Replicated Data Type). This approach consists of modeling a document as a sequence of items, each with a global unique identi er, and providing insert/delete operations on those items. The dynamic generation of these identi ers is a challenging problem, since limited consumption of resources and reliable propagation of identi ers must be guaranteed. The paper presents a novel strategy for such a generation that works well with a large number of users and has a very limited impact on latency.
In Tracking Changes Through EARMARK: a Theoretical Perspective and an Implementation the authors deal with changes on markup structures. Their goal is to de ne in a precise and unambiguous way when (and how) the same markup element has to be considered as changed, if its content changes. The authors propose a theoretical representation of change tracking information based on FRBR (Functional Requirements for Bibliographic Records), that also provides support for expressing provenance information. Their implementation of the framework is based on EARMARK, a Semantic Web-based meta-model that allows a ne-grained de nition of overlapping structures on plain content.
The workshop also includes a round-table session. The outcome of the discussion is not reported in these proceedings, since the workshop is not yet held. Some topics that will be discussed are: distributed editing issues (following-up the keynote), human-interpretation of changes and quality of deltas, and identi cation of editing patterns in other domains like law-making and humanities. More updated details will be published on the workshop web page http://diff.cs.unibo.it/dchanges2013/roundtable/. Suggestions from the audience will be encouraged throughout the workshop. Our goal is to foster research collaboration and to also identify topics for a second edition. We hope to have more and more interesting editions of DChanges in the future and to gather a lively community of researchers around these themes.
In conclusion, we would like to thank all people who had expressed interest in DChanges and the organizers of DocEng { in the rst place Simone Marinai and Kim Marriott { for giving us the possibility of organizing it and for supporting us continuously.
Our thanks go to the committee members, for their hard work in circulating the call for papers and reviewing papers (perfectly on time!).
A special thank goes to Ethan Munson, for his illuminating keynote.
The workshop has been organized by four people, from two research groups. They have been helped by a committee of experts from all around the world.
Gioele Barabucci is a research fellow at Universita di Bologna. He recently received his PhD with a thesis on di algorithms and delta models.
Uwe M. Borgho is a full professor of Computer Science at Universitat der Bundeswehr Munchen. With his research group, he published various papers on algorithms for comparing textual documents and on related topics.
Angelo Di Iorio is an assistant professor at Universita di Bologna. He worked on various systems for document versioning and publishing, and collaborative editing.
Sonja Maier is a Postdoc at Universitat der Bundeswehr Munchen. In her research, she focuses on tool creation and tool integration for (visual) domain-speci c languages, and is interested in tracking the evolution of text and diagrams.
Serge Autexier, DFKI Bremen Stephane Ducasse, INRIA Lille Nord Europe research center
Yannis Tzitzikas, University of Crete and FORTH-ICS
Merging Uncertain Multi-Version XML Documents
Mouhamadou Lamine Ba, Talel Abdessalem and Pierre Senellart Identifying Change Patterns in Software History
Jason Dagit and Mathew Sottile The Concept Di erence for EL-Terminologies using Hypergraphs
Andreas Ecke, Michel Ludwig and Dirk Walther An Algorithm for Transforming XPath Expressions According to Schema Evolution
Kazuma Hasegawa, Kosetsu Ikeda and Nobutaka Suzuki Tracking changes Through EARMARK: a Theoretical Perspective and an Implementation
Silvio Peroni, Francesco Poggi and Fabio Vitali