Me4MAP V1.0: A method for the development of metadata application profiles

Tracking #: 1776-2988

Authors: 
Mariana Curado Malta
Ana Alice Baptista

Responsible editor: 
Eero Hyvonen

Submission type: 
Full Paper
Abstract: 
This article presents Me4MAP V1.0, the first version of a method for the development of Metadata Application Profiles (MAP). A MAP is a construct of the semantic web (SW) that enhances interoperability. It is a data model that provides reference information about the global context of both the related data and its constraints. The development of Me4MAP was based on a Design Science Research methodological approach and has as starting points the Singapore Framework for Dublin Core Application Profiles and the Rational Unified Process. Me4MAP integrates knowledge from: (i) the practices of the metadata community concerning MAP development, and (ii) software development processes and techniques, focusing on the early stages of the processes that deal with data modeling. Me4MAP establishes a well defined process for the development of a MAP through definition of activities, when those activities should take place, how they interconnect, and their resulting deliverables. Me4MAP also suggests techniques to be used in building the deliverables. By making Me4MAP available to the SW community, we hope to contribute to the enhancement of MAP development.
Full PDF Version: 
Tags: 
Reviewed

Decision/Status: 
Reject

Solicited Reviews:
Click to Expand/Collapse
Review #1
By Marcia Zeng submitted on 09/Apr/2018
Suggestion:
Major Revision
Review Comment:

This manuscript was submitted as 'full paper' and should be reviewed along the usual dimensions for research contributions which include (1) originality, (2) significance of the results, and (3) quality of writing.

The paper “presents Me4MAP V1.0, the first version of a method for the development of Metadata Application Profiles (MAP).” The project reported is a solid research and has been in development with the collaboration of high level researchers in the metadata field for the past several years. It uses five sections to briefly present: 1&2: “the DSR methodology as well as the application of the Hevner’s framework.” 3: “the method for the development of MAPs.” 4: the limitations of the work. 5: conclusions and future work.

A clear roadmap of developing a MAP would provide the benefit to the society which is embracing more and deeper development of structured data in the Semantic Web era.
However, the paper is not easy to read or follow through by people who are not in the metadata field, or even narrower, in the Dublin Core related metadata field.

If the paper would like to reach a broader community, showing a set of paths of developing metadata application profiles following the method used by Me4MPA, then the paper needs to be written with more focused delivery. Right now, it is still like an internal report about how Me4MAP was developed. Even though this is a helpful and meaningful topic, it should consider the readers and developers outside of the circle. Therefore, the first suggestion is to modify and deliver very clear research findings.

As a research project, the paper did not include a review of literature or related development. It is not clear how many, in the world, similar approaches have been explored or existed and how realistic this Me4MPA is today.

There is also no report about if there is any testing involved and refinements conducted I saw, based on related publications by the authors, that there were implemented cases in its earlier draft. Here the paper reported some discussions at workshops, but what were the impact to this project?

The paper needs to bring the contextual information to the text, especially the beginning sections. Keep giving people the reference numbers instead of any author names and title or key concepts is not a good way to lead the reading. (See this paragraph, something ‘developed by [12]’, “according to [9]”, etc.)
“The version presented here uses a Design Science Research
(DSR) methodology that follows the framework developed
by [12].
According to [9] a MAP is a very important construct
to achieve interoperability.”
It is also inconsistent when referring to citations, where the needed citations were missed. What is “the Hevner’s framework” when it was brought in the first time?
“The following section briefly presents the DSR methodology as
well as the application of the Hevner’s framework.”

The text needs to correspond with the figures. You said S1 –S4 (stages). It will be necessary that Figure 2 shows such codes. The mix use of A (e.g., A1) and S (e.g., S1) at the same level of sections in the paper does not help, only causes confusion.
“ 3.4. A2 – Development of the MAP Documentation
3.5. S1 – Developing the Functional Requirements
The third Singapore Stage (S3) is made of four subactivities
(Figure 7):
– S3.1 - Definition of the Vocabulary Alignment (see Table 13);
– S3.2 - Definition of the Constraints Matrix (see Table 15);
– S3.3 - Constraints Matrix Test (composite activity - see Figure 9);
– S3.4 - Encoding of the Description Set Profile - Short Name “DSP Encoding” (see Table 19).”
The figures and tables following this statement (p.8) are far away from what was referred (p.10 and later).

Overall, the delivery of this report is not easy to follow without spending much time to figure out all the codes, tables, and figures (in addition to the non-familiar concepts and terms). Maybe the layout of the pages added inconvenient of the reading.

Review #2
By Koraljka Golub submitted on 10/Apr/2018
Suggestion:
Minor Revision
Review Comment:

This manuscript was submitted as 'full paper' and should be reviewed along the usual dimensions for research contributions which include (1) originality, (2) significance of the results, and (3) quality of writing.

The paper is a valuable contribution towards improving the quality of metadata schemas in general and application profiles specifically. The proposed framework is theoretical at this stage, and a follow-up paper with tested methodology would be an important complement.

Very minor corrections are needed such as aligning capitalisation of headings, usage of - versus -- etc.

Review #3
Anonymous submitted on 14/May/2018
Suggestion:
Reject
Review Comment:

The authors present a methodology Me4MAP for creating Dublin Core metadata application profiles (MAP) widely used on the web and in linked data applications. The contribution of the work is based on the fact that there are only very vague guidelines, such as the Singapore Framework developed by the DC community, for creating MAPs. The topic is suitable for SWJ.

The paper well-structured, well-finished, and written in good English.

In section 1, a nice motivating introduction to the topic with references to DC documentation and literature is given.

After this in section 2, the methodology used in designing Me4MAP itself is described based on principles of Design Science Research. Although the process is documented in detail with several references I found this section not very useful as it is: the underlying ideas cannot easily be understood without having read the underlying works, such as Hevner's three cycle framework and Rational Unified Process (RUP). These works are probably not well known to the readers of SWJ and more explanation is needed here. It is a good idea to use some methodology in designing Me4MAP but the point of this section from the perspective of the paper's own contribution, i.e., the Me4MAP method itself, should be clarified. Tell the reader what is your message and lessons learned here in addition to just documenting a process in detail.

In section 3, a very detailed description of the methodology is presented including its stages, activities chart, development team description, and over 20 tables of activity descriptions. However, I have concerns related to the presentation and results of the work at its current stage. First, although the descriptions are detailed they are very abstract and no examples of actually applying the method to some use cases are presented. Clarify what is the problem your are actually solving here when developing MAPs. Based on the presentation, it would be difficult to actually use to the methodology to creating a new MAP. Second, it is not explained why these particular activities and the project team composition was selected but the model presented rather as a given truth, based on detailing out the Singapore Framework. This may be a natural and useful step ahead but still some kind of discussion, motivation, and evaluation of the model is needed in order to convince the reader that Me4MAP really is "good" and worth using.

The paper concludes with a short section focusing on some limitations of the work and directions for future work. The authors state that this work is not a universal solution but a first step towards methodology for developing better application profiles. Also in my mind, such methodology would be useful but needs to be validated with real use cases showing its benefits and possible challenges before publishing it as a SWJ journal article. The paper would require more fundamental additions and changes than just revising the current manuscript, and therefore I recommend rejecting the paper at this stage, even if the papers also has virtues and potential for further development.