3 WWW and Java-based system for investigation of mapping algorithms 5 CEV: Collaborative Environment for Visualization Using Java-RMI 11 Adaptive Placement of Parallel Java Agents in a Scalable Computing Cluster 14 Cracking RC5 with Java applets 19 JSPICE: A Component-based Distributed Java Front-End For SPICE 21 Fornax: Web-based Distributed Discrete Event Simulation in Java 22 Multi-User 3D Scene Model Editor in Java 27 GVis: Interactive Visualisation of Genetic Algorithms 28 A Java Framework for Seamless Sequential, Multi-threaded, and Distributed Programming 40 JWarp: A Java Library For Parallel Discrete-Event Simulations 42 Towards Bayanihan: Building an Extensible Framework for Volunteer Computing Using Java NOT AVAILABLE when I left on trip ...... 46 JAVA Agents for Distributed System Management 56 Algebraic Java classes for numerical optimization ------------------------------------------------------------------- 1998 Java Workshop Referee Report --------------------------------- Please return this report by January 8, 1998 to java98@cs.ucsb.edu Paper Number: Paper Title: Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: Quality of presentation: Interest/Importance to current workshop: Confidence in your evaluation (overall): Recommended action for paper: (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) Regardless of the recommendations, please write specific comments and suggestions that can be communicated to the author(s). The first paragraph should be a summary of the paper, and an enumeration of the new ideas/approaches/results claimed by the authors. The rest of your review should be a detailed critique of these claims, including discussion of the following questions. Are the claims of novelty justified? If so, are the proposed techniques practical? If they are practical, are there experimental results to show that it is worth implementing them? Are the references adequate? ------------------------------------------------------------------- Confidential Portion of Report ------------------------------ Is there anything about the paper that you would like to communicate to the program committee in confidence? Referee name: Referee affiliation: Referee email address: Program Committee Member who assigned the paper to you: (for non-PC referees only) ------------------------------------------------------------------- -------------------------------------------------------------------- Klaus E. Schauser E-mail: schauser@cs.ucsb.edu Assistant Professor Phone: (805) 893-3926 Department of Computer Science (805) 893-4321 University of California FAX: (805) 893-8553 Santa Barbara, CA 93106 http://www.cs.ucsb.edu/~schauser -------------------------------------------------------------------- 1998 Java Workshop Referee Report --------------------------------- Please return this report by January 8, 1998 to java98@cs.ucsb.edu Paper Number:3 Paper Title:WWW and Java based ... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper:2 Quality of presentation:3 Interest/Importance to current workshop:2 Confidence in your evaluation (overall):4 Recommended action for paper:2 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- The VIM approach is controversial but this is not why I marked paper down. I thought VIM was irrelevant to paper, which described a useful but not very deep applet. END PAPER 3 **************************************************** Paper Number: 5 Paper Title: CEV Collaborative .... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 2 Quality of presentation: 3 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 5 Recommended action for paper: 2.5 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- The authors appear not to have fully read the literature, which they mention in sec 1.2 but in spite of a claim to the contrary, do not compare with their work in the rest of paper. I am quite knowledgeable about this field but could not see which feature of CEV, authors viewed as novel. Comparative systems are Habanero (NCSA), Tango (NPAC) and SciVis (NPAC again). I see CEV as a project in an interesting area but this paper is a little premature and the authors need to develop the distinctive features of their approach and as the system matures, naturally they will also improve the evaluation and deductions which are currently skimpy. END PAPER 5 **************************************************** Paper Number: 11 Paper Title: Adaptive Placement .... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 3 Quality of presentation: 3 Interest/Importance to current workshop: 3 Confidence in your evaluation (overall): 4 Recommended action for paper: 3 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- This paper addresses an interesting topic but is disappointing (to me) as it focuses on load balancing (which is well studied and not much to do with Java) and the only real application is Jacobi Iteration for which we undoubtedly do not need agents. I am also surprised that an agent paper would focus on small clusters as what I think of as interesting agents are at their most important in "world wide" heterogeneous systems. END PAPER 11 **************************************************** Paper Number: 14 Paper Title: Cracking RC5 .... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 3 Quality of presentation: 3 Interest/Importance to current workshop: 3 Confidence in your evaluation (overall): 5 Recommended action for paper: 3 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- There are other similar Web approaches to cryptography and this paper uses nicer technology but is otherwise less sophisticated than some of previous efforts. For instance earlier work used multiple servers and a more asynchronous fault tolerant approach to doling out work. I believe these features could and should be added to current approach. So although I think this paper describes slightly undeveloped ideas, I did like their precise comments on the modest performance shown by the Java code. END PAPER 14 **************************************************** Paper Number: 19 Paper Title: JSpice Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 3 Quality of presentation: 4 Interest/Importance to current workshop: 5 Confidence in your evaluation (overall): 4 Recommended action for paper: 4 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- This is an example of an important class of the application of Javabeans for packaging (legacy) codes. Some comments on the 100% OMG (as opposed to 100% Java) CORBA view of this example would be helpful, as there is some expectation that Javabeans will become component (visual) front end to CORBA. Some people have complained to me about performance problems with Java Studio. I look forward in general to the more detailed discussion, which will be very important. ----------------- Comments to Program Committee ------------------------- An important paper in my opinion as it is an exemplar of a process that I think will become very common with CORBA and Javabeans being used as wrappers for either distributed or local execution of codes. END PAPER 19**************************************************** Paper Number: 21 Paper Title: Fornax Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 4 Quality of presentation: 4 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 4 Recommended action for paper: 4 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- A nice discussion of event driven simulation with occasional English glitches (save not safe, form not from in sec 2.2; critical sentence at end of sec. 5.5 is nonsense I think). A discussion of the US military standards HLA/RTI (from DMSO in US but adopted internationally) would be helpful. ----------------- Comments to Program Committee (same for 21,40,57) ------------------------- I read papers 21,40 and 57 together as all address the same general issues in parallel (a.k.a. distributed) discrete event simulation. I have always thought that this is very natural and important for Java. Surprisingly (to me) no one mentioned HLA/RTI, which I understand, is becoming an international CORBA like specification for military distributed event driven simulations None of the papers had made great progress but all had some results. I was not certain that papers actually "agreed" on issues. I did not find great differences in quality but I rated them n order 21(top) then 57 then 40. I don't think we should have more than one presentation at conference but could ask perhaps for one speaker to discuss all 3 with an emphasis on their work. END PAPER 21 **************************************************** Paper Number: 22 Paper Title: Multi User 3D Scene ...... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 2.5 Quality of presentation: 3 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 4 Recommended action for paper: 3.5 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- I believe contains a lot of interesting ideas in two important topic areas: a) Shared Document Preparation b) Shared Construction of a 3D Virtual World I believe that the authors reduce b) to a) by noting that VRML is an ASCII stream. However I would have thought that the 3D world would have been more naturally edited using a higher lever visual representation which would also have interesting sharing issues. The shared text(document) editor is such an important real world product that it needs to be considered in general rather than for special case of VRML. END PAPER 22 **************************************************** Paper Number: 27 Paper Title: Gvis: Interactive Visualization .... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 2 Quality of presentation: 2 Interest/Importance to current workshop: 3 Confidence in your evaluation (overall): 4 Recommended action for paper: 2.5 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- This is a workmanlike paper developing a Java front end (and not really a sophisticated visualization system) for genetic algorithms. There do not appear to be many issues that are specific to the particular application area -- Genetic algorithm based optimization Thus paper would be stronger if placed in a more general context e.g. general optimization front-end or general scientific visualization. There are no references to the many relevant papers (which are not necessarily Java based) on client-server visualization. END PAPER **************************************************** Paper Number: 28 Paper Title: A Java Framework for Seamless ..... Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 4 Quality of presentation: 3 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 3 Recommended action for paper: 4 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- This paper would benefit by tying its ideas to existing successful parallel computing models. Further it needs to be a little more up front about the difficulties with their model and what application class it is aimed at. There are some obvious performance problems with this approach and it is not clear to me if these are assumed to disappear in future implementations or to restrict the application class addressed by this environment. I very much doubt if matrix problem looked at in paper, is a compelling motivation for this approach. I wonder if Java// is a possible backbone for the parallel discrete event simulation presented in other papers at this workshop. END PAPER **************************************************** Paper Number: 40 Paper Title: JWarp Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 3 Quality of presentation: 3 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 4 Recommended action for paper: 3.5 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- I would suggest discussion of distributed memory DES -- these are dominant DES implementations in the U.S. although in some cases it is true that event driven simulations require the low latency of shared memory and hence can be implemented in a single Java VM. I would also like to see an evaluation of the adequacy of the Java thread model and Java object serialization (for rollback) for time warp. ----------------- Comments to Program Committee (same for 21,40,57) ------------------------- I read papers 21,40 and 57 together as all address the same general issues in parallel (a.k.a. distributed) discrete event simulation. I have always thought that this is very natural and important for Java. Surprisingly (to me) no one mentioned HLA/RTI, which I understand, is becoming an international CORBA like specification for military distributed event driven simulations None of the papers had made great progress but all had some results. I was not certain that papers actually "agreed" on issues. I did not find great differences in quality but I rated them n order 21(top) then 57 then 40. I don't think we should have more than one presentation at conference but could ask perhaps for one speaker to discuss all 3 with an emphasis on their work. END PAPER 40 **************************************************** Paper Number: 46 Paper Title: JAVA Agents Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 4 Quality of presentation: 5 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 4 Recommended action for paper: 4.5 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- An excellent paper. Some more details on performance (speed) and effectiveness (value to system) would be useful END PAPER **************************************************** Paper Number: 56 Paper Title: Algebraic Java Classes Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 4 Quality of presentation: 4 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 4 Recommended action for paper: 4.5 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- An excellent original paper. I would suggest more detail on the desire for [] overloading. I am not certain that I understood this correctly. END PAPER **************************************************** Paper Number: 57 Paper Title: JTED Place a number corresponding to your evaluation for each item below. (5=outstanding, 4=high, 3=medium, 2=fair, 1=poor) Technical contribution of paper: 4 Quality of presentation: 2 Interest/Importance to current workshop: 4 Confidence in your evaluation (overall): 4 Recommended action for paper: 4 (5=strong accept (award quality), 4=accept (I learnt something), 3=maybe (if we are hard up), 2=reject (will raise eyebrows if accepted), 1=strong reject (will be an embarrassment if accepted)) ----------------- Comments to Author ------------------------- I found the colloquial style unnecessary and detracting from the message of the paper. A discussion of possible impact/relevance of the new DMSO HLA/RTI would be interesting. I did not understand the seemingly important discussion of which processes could be threads. To be colloquial, either improve discussion or require more intelligent referees ...... ----------------- Comments to Program Committee (same for 21,40,57) ------------------------- I read papers 21,40 and 57 together as all address the same general issues in parallel (a.k.a. distributed) discrete event simulation. I have always thought that this is very natural and important for Java. Surprisingly (to me) no one mentioned HLA/RTI, which I understand, is becoming an international CORBA like specification for military distributed event driven simulations None of the papers had made great progress but all had some results. I was not certain that papers actually "agreed" on issues. I did not find great differences in quality but I rated them n order 21(top) then 57 then 40. I don't think we should have more than one presentation at conference but could ask perhaps for one speaker to discuss all 3 with an emphasis on their work. END PAPER 57 ****************************************************