The paper is very interesting and brings up a lot of issues. It is relevant to t hose who are interested in Internet security and how OS's provide security for o r against agents. However, the language in certain sections of paper are too casual such as using terms as "bad guys." One last note, try not to use corporate trademarks or names such as "Delta Airlines" in your examples. There are some examples in the paper to make some of the ideas more lucid. The p aper covers a wide breadth of topics about agents and their host operating syste ms too. However, the tone of the paper in general sounds like it wants to avoid giving solutions to problems it presents. As said in its conclusion, this area o f research is relatively premature and requires much more development. It almost sounds like a bias of some sort and a way to dodge indepth analysis of certain topics covered in the paper. The presentation of the solutions and their criticisms went hand-in-hand. Howeve r, I had a hard time figuring out the abstract mechanisms on what these solution s are based upon. ====================================================================== The authors present security challenges faced by Mobile Code Systems in this paper. Being a relatively new concept, some aspects of security have not matured yet. Hence, as pointed out in the paper research efforts in this field should converge on certain techniques to gain wide spread usefulness. Certain points that I felt were not clear are: - Proof-carrying is a technique used to prevent host corruption, where the client attaches a proof of adherence to its' publicized security policy. However, what happens if it attaches a bogus proof? Not cross examining the code and the proof might prevent the host from catching disparities between the code and the proof. - When hardware is provided to take over the proof production and verification in proof delegation, wouldn't that piece of hardware become a bottleneck? Replication might be too expensive a solution, provided that such hardware is not widely available. In conclusion, I really think this was an interesting paper. It successfully summarizes the issues related to mobile code security. However, occasional use of informal language, in my opinion, take away from the overall quality of the paper. ====================================================================== The paper has a well defined outline, and explains clearly the security problems that result from using mobile agents. Even though the topic is "hot", as are many Internet-related topics, the authors looked at one of the most important current issues in computing - security, at the difficulties of implementing it, and have analyzed and summarized these issues well. There is a good presentation of the problems that arise with the use of mobil agents, and the techniques of dealing with these problems. There are examples that support the claims and observations. The research is interesting, and novel in many ways, it looks at the latest development in a relatively new and not well defined area. The authors emphasized the questions that remain unanswered in the the filed, and pointed out a few possible directions that a future research might take. Overall the paper deserves an excellent score. ====================================================================== IMPORT: In my view, the subject area discussed is quite an emerging one and of importance, especially in the light of the growing popularity of Internet. This paper does well in picking up the security concerns associated with both the host and agent security, which we are going to face in time to come. NOVELTY: The paper presents some future avenues and poses some open questions towards the end, which lend some thoughtfulness to the subject. QUALITY: The authors have clearly delineated the various security issues dealing with the mobile code systems. Although, at many places, the ideas presented looked abstract to me, but in general, I would say that the observations made have been supported well. OVERALL: The overall structure of the paper is very good and clear in its scope. However, I think if the paper can provide a little bit insight into possible implementation aspects about the various security policies discussed, it would add to the value of the paper. ====================================================================== The paper gives a good overview of mobile code and the concept of the mobile agent computing paradigm. Overall, the paper lacks analysis contents. The paper raised many interesting questions on the design and the approach to take. However, the author does not provide enough guidance as to determine the viability of the technologies presented. ====================================================================== I do not believe that the paper effectively tied itself to the topic of operating systems. The Subject of the paper was fairly novel and the paper was well written. I believe that in a few cases the text dropped below formality expected by the intended audience. Some of the adjectives used and the style used are better suited for conversations. For example, the acknowledgments section, horrendous bungles, and TaekYorDo individually add some humor to the paper perhaps there was too much humor. ====================================================================== 2, Survey of security in Mobile Code Systems(SS) Import Security is a big problem in modern computer design, mobile code system is a very new idea and "hot" topic in OS. The authors select security in mobile code system as their survey,it can attract many computer researcher's attention and it is an excellent survey topic. Novelty Although many ideas are from reference papers, the authors give a reasonable analysis, and put related information together(such as 4 Host security, 5 Agent security and the conclusion ). This gives readers some new ideas about how the security problems were solved in different clients and servers. Quality This paper has good quality. The authors provide different techniques employed to ensure host integrity and agent security. Besides they use Java applets and ActiveX controls as examples to make readers understand their topic easily. Overall This is an excellent survey paper. First I was impressed by their topic.I'd like to read it to know more about mobile agents, and relative security problems. Second, the paper has a good written structure, so the general idea of this paper is easy to be understood, although some techniques are very complicated and many of them are still theoretical. Overall, we should accept this paper. (P.S.)I'd like to know more mobile agent examples except Java applets and ActiveX. ======================================================================