ESWC Review

=GoodRelations Rejection Letter at ESWC 2008 =

The first attempt to publish a paper on GoodRelations at an academic conference was a disaster ;-). Basically, the reviewers complained that the ontology was not complicated enough, and that building an ontology was no scientific contribution to the Semantic Web.

You may want to have a look at the papers that were actually accepted ;-).

Here is the original letter.

Subject: ESWC 08 notification From: eswc08@easychair.org Date: Fri, 29 Feb 2008 16:59:12 UT To: mhepp@computer.org

Dear authors,

The ESWC Programme Committee has now finished the task of selecting papers for the European Semantic Web Conference 2008.

We regret to inform you that your paper has not been accepted for publication. Competition for ESWC this year was as fierce as ever and selecting the programme was challenging. Each paper was reviewed by three independent reviewers, with a Senior PC Member overseeing the process for each paper. Papers were then also discussed at a Senior PC meeting which was responsible for the final decision. We received over 270 submissions, of which only 51 were finally selected for publication.

We enclose the reviews of your paper. Please note that the call for poster submissions is open until March 14th, so you may wish to consider submitting a poster about your work.

We thank you for submitting your paper to ESWC'08 and still hope to see you in Tenerife in June.

Yours,

Sean Bechhofer Manolis Koubarakis Programme Co-Chairs, ESWC'08

-

Paper: 139 Title: GoodRelations: An Ontology for E-Commerce on the Semantic Web

review 1 OVERALL RATING: 1 (weak accept) REVIEWER'S CONFIDENCE: 2 (medium) Candidate for Best Paper?: 1 (No) --- REVIEW

Three Sentence summary of your review:

This paper develops an ontology for representing knowledge about commercial offerings. The work is not deep as such but it is quite useful. It will lead to good discussions at the conference.

Detailed Review:

This paper addresses the problem of sharing information about commerce regarding commodity products and services. It proposes the GoodRelations ontology for describing such commercial offerings.

The author has made an effort to capture knowledge of this important domain. The contribution is apparently about how well this knowledge is organized in this ontology. The ontology proposed seems reasonable enough. The various parts of it are drawn from other works, which lends some credibility to this work. However, there is no evaluation as such (the author notes that an evaluation is going on).

I expected to see a comparison with existing efforts such as the Universal Business Language and more importantly the UNSPSC. It seems from the author's other publications that he has studied such systems before.

The generic ideas of this paper dealing with ontology representation seem to have been published already. If not, the author should clarify differences from previous works, especially his own.

Overall, the work is useful and possibly interesting to the conference attendees, but not very deep.

Figure 1 is illegible without considerable magnification. I couldn't read it from the hardcopy of this paper.

review 2 OVERALL RATING: -1 (weak reject) REVIEWER'S CONFIDENCE: 3 (high) Candidate for Best Paper?: 1 (No) --- REVIEW

Three Sentence summary of your review: This paper presents a description of an ontology about Web offerings in e-commerce, which is aligned with existing ontologies used for the description of products and services (e.g., eClassOWL).

Detailed Review: The paper describes an ontology that is the result of a long effort (started in 2005) to describe offerings of products and services in the Web. The author claims at the beginning that this ontology should be a non-toy ontology (which is the case in most of the current state of the art in this area), which is able to describe aspects related to e-commerce such as Web resources, offerings, legal entities, prices, and terms and conditions. This is quite an ambitious objective, specially in the areas of offerings and terms and conditions, given the large amount of possibilities and legal aspects involved in this area.

While this is an effort worth taking into consideration, the feeling after reading the paper is that there are many limitations to both the end result and to the process followed to reach the end result, together with a lack of valuable lessons learned that motivate my decision about this paper: - The areas of offerings and terms and conditions are only dealt with quite superficially, that is, even though the author claims that this ontology is a non-toy one, my impression is that the complex areas are weakly-considered: only warranty is considered in the ontology and is not too complex. How do you deal with aspects related to different countries? And what if the warranty is more complex than the three types that you describe? What about forms of returning goods? These are only some of the questions that are not covered by the ontology. - This can be deducted as well from the list of competency questions provided in section 2. It is not clear whether this list is exhaustive or not, but definitely it covers only a few aspects involved in the general area of e-commerce. - Section 4 could be considered initially as a "lessons learned" section, which would be of high interest to the audience of this conference. However, the author does not provide many new lessons learned or a good classification of why these lessons learned are unique or at least very relevant for the e-commerce domain and not for other domains, or how they can be extended to other domains. Examples such as how to workaround n-ary relations in OWL, or how to deal with ranges, or how to distinguish between two types of products are common in the knowledge engineering literature, and do not provide any additional insight. - The same applies to the discussion about URIs. First, there is a mistake when the author talks about the fact that Web resources have a URI assigned. I think that the author is mistaking URI (the identifier) by URL (the locator). Anybody can say anything about a URI even if this is not under the domain of control of such person. Besides, I think that the author should really propose a model for naming resources in the e-commerce domain based on his experience. This is something that has been done, for example, in other domains such as biology (with the use of LSIDs). - The evaluation section does not provide much about how the evaluation has been carried out and about the exhaustiveness of such evaluation. In theory, the ontology should be contrasted with the competency questions that were useful to generate the ontology requirement specification. However, the author only shows a few SPARQL queries. This is not a real evaluation.

Finally, as a summary, the author does not provide much insight either in how to model e-commerce relationships in general (the model is just a small, toy-like, model) or in how to apply ontology development methodologies in such a long and complex process, so I hesitate about whether this paper can provide much benefit to the conference attendees, although at the same time I acknowledge that this model may become a good baseline for representing e-commerce relations in the Semantic Web.

review 3 OVERALL RATING: 1 (weak accept) REVIEWER'S CONFIDENCE: 2 (medium) Candidate for Best Paper?: 1 (No) --- REVIEW

Three Sentence summary of your review:

The paper presents an ontology for the semantical annotation of web offerings. It also discusses how this ontology can be represented in an light weighted approach.

Detailed Review:

The paper is well written, a little bit difficult to understand, but I like that the author also argue for his decisions. The only what I have to remark is:

- please correct the English mistakes including "be a be a" (page 6) and "evaluatiing" (page 13)

- revise your writing style in order to adopt the more objective research paper style. Not everyone prefers to be entertained.