I enclose 3 Referee reports on your paper. We would be pleased to accept it and could you please send me a new version before November 5 99 Please send a memo describing any suggestions of the referees that you did not address Ignore any aggressive remarks you don't think appropriate but please tell me. I trust you! Thank you for your help in writing and refereeing papers! Referee 1 ******************************************************** Subject: C430 JGSI review Paper: C430 Title: A Java/CORBA based Visual Program Composition Environment for PSEs Overall Recommendation: Weak-Reject This paper describes a Visual Programming Composition Environment (VPSE). While the paper does a good job of describing the system design and is well written, in my view it does not sufficiently describe the goals of the work being described in the paper or the research contributions. Especially with respect to the topic of the special issue "Java Grande". For this venue I think an experience paper showing how this system was used to solve Java Grande types of problems would be more appropriate. As currently written too much of the paper concentrates on the system architecture and as a result the paper seems more suited to a journal on software engineering or system architecture. According to Section 5 (Conclusion) "... The difference in our approach is the pervasive use of XML throughout the system, ..." This does not seem like a significant research contribution - unless a clear and precise explanation of the advantages and disadvantages of this approach are given. The section on Previous Work does describe previous work and the differences from the work in the paper, however, I think the paper could be significantly improved by taking this section a step further and explaining the advantages and disadvantages of these differences. E.g, when comparing with Gateway the use of XML is touted as a difference. What are the implications of this difference? The authors of the paper should share the insights gained in working on this project with the reader. The prototype VPCE, an "elementary mathematical equation editor" doesn't really apply to the problem of trying to use Java and several or large machines to solve grand challenge types of problems. A more in-depth comparison with JavaBeans would be useful. How is this system significantly different from JavaBeans? What are the advantages and disadvantages of this approach versus JavaBeans? It seems that a statement in abstract and later in the paper are contradictory - but after reading it a few times I think I understand. Maybe the abstract could be reworded to make it more clear. >From the abstract: "... compiling running applications in a *specific problem area of domain*". >From the first line in the Introduction (and elsewhere in the paper): "A CORBA-based *domain-independent problem solving environment ..." Referee 2 **************************************************************** Subject: C430 JGSI Review Resent-Date: Mon, 11 Oct 1999 08:25:07 -0400 Resent-From: Geoffrey Fox Resent-To: p_gcf@npac.syr.edu Date: Thu, 7 Oct 1999 11:50:29 -0500 From: Joel Jones To: Geoffrey Fox >a)Overall Recommendation Excellent paper, I highly recommend it for publication. >b)Words suitable for authors There is little to say about your paper. It is to the point, describing the problem it solves, other approaches, etc. The top of page 5 is particularly good. The second paragraph on page 11 could be worded a little more clearly, where it enumerates the two approaches to event handling. I would either eliminate figures 2-4, or replace them with line drawings. A page and a half is a lot of space to use to present so little information. Similar comments hold for figures 5 and 6. It might be beneficial to show the dialog of figure 5 as just a text figure, rather than showing all the buttons and window decorations. >c)Words for me if necessary Referee 3 ************************************************************** Subject: C430 JGSI Review a) publish b) This paper describes a modern multi-tier system to assist the user to construct applications from distributed components, and may be used to construct a problem solving environment (PSE). It is entirely implemented in Java, and uses XML to define all necessary interfaces. Unfortunatelly, the authors do not define what is a component, nor they describe what component model they are following. From examples they provide in the text, one can deduce that they use JavaBeans approach somehow "upgraded" to fit distributed environment. In the title of the paper, they claim to use CORBA for inter-JVM communications, but never discuss it in the text. The neural-network example uses Java-RMI instead. This is definitely a flaw in the presentation of this interesting work. Also, Figure 1 one is hard to interpret: it is not clear which elements of the system comprise a front-end, and what is implemented in middle tier(s). It is not clear to me, whether IRMS ("intelligent resource management system") is a part of VPCE (subject of this paper) or not. Definitely, a natural place for IRMS is the middle tier. Nevertheless, I found the paper very interesting, in particular the idea of a repository of component interfaces specified in XML (with embedded documentation), component user interfaces defined in XML (to be rendered by the front end), and describing the final application as a connectivity graph expressed in XML. All together, it promisses flexibilty and extensibility of the system, and to large extend automated generation of the user front-end allowing the user to visually compose applications from predefined components, including legacy applications wrapped as components. c) this paper is hard to read. The XML specifications are not complete, many aspects are missing. For example, discussion of events it out of context. It happened that I talked to Omar in San Francisco but for unprepared reader it may be confusing.