This paper presents XCDL, a visual language based on colored Petri nets that represents Extensible Markup Language (XML) documents and supports XML operations and manipulations. The proposed language also supports querying and data manipulations such as insertion and deletion.

The authors claim that existing tools focus on query support. The novelty of the proposed tool is that it supports both querying and data manipulation; it’s claimed to be superior for this reason. The paper begins with both casual and formal definitions of the language, and continues with the results of a performance evaluation.

Both the topic and the presentation are interesting, and the paper flows well. However, in many places, it may be hard to understand the concepts or follow the discussion. Figures are scattered throughout the paper, and are referenced either pages before or after the figure appears in the text.

This work presents a new tool and a new language, intended for both novice and experienced users. However, to understand the language, the reader must have some background in Petri nets and XML, as well as SQL and other querying languages.

The evaluation of the tool and the language is very weak. Proper empirical evaluation would give further evidence of support for this tool and language. The paper could have been significantly enhanced if a single motivating example was used to explore the language. This would have greatly helped readers understand the tool.

The evaluation section is surprising. First, it is not an evaluation and so the title of the section is inaccurate. Evaluation, at best, has been left for future work. But even if you grant the decision to call it an evaluation, the paragraph is weakly written and obscure. The authors refer to a positive feedback of 87 percent for their tool, but less than 87 percent for other tools. What does this mean, and what does “less than 87 percent” mean? Is it 86 percent, or 45 percent? Also, how many participants were involved? What is considered positive and negative? What statistical testing was used? The evaluation should support the claims made earlier in the paper, particularly the claims of simplicity, expressiveness, flexibility, scalability, and adaptability.

On page 112, mashups need to be explained sufficiently so that the reader can understand why the scenario cannot be solved using existing mashups or XML visual languages. The authors note: “To the best of our knowledge, no tool yet provides information regarding the analysis of the performances.” Performance analysis for what exactly? When and why is performance analysis useful? Perhaps no tools can accomplish performance analysis simply because it is not needed. The motivating example does not uncover the need for performance analysis.

In section 3.2, it would have helped readers to see the distinction between a visual query language and a visual representation for the data itself, using XML in this case. These are totally different, but this paragraph mixes them up and it is not clear which parts of the discussion are about which concept.

Figure 1 and table 1 are not properly explained in the text. Section 4 refers to models as pictures. My view is that a visual representation of XML data is actually a model with modeling elements (such as nodes and links). Finally, section 5 does not justify the claim that the language is user friendly if it has the five properties mentioned in this section. User evaluation would have helped to support such a claim. Also, without empirical evaluation, we have no way to determine whether the language even has the five properties.