------------------------------------------------------------------------------ This paper packed a lot of good information in fourteen pages. This is the reason for the 5 rating in Import and even higher rating in Quality. I agreed with most of the authors' observations; though most seemed like observations that any resonable person would make after reading the same reference papers (which lead to the below average 3 in Novelty), they were sound and clearly written. With my Overall rating of 6, I think this paper should be accepted into the Proceedings of OSSurveyF98. ------------------------------------------------------------------------------ Paper provided a good summary of the different approaches and had some good observations/criticisms. Seemed like a well written paper with most concepts well defined so as to make it easier for the user to read & comprehend. The one thing was that the approaches discussed were assumed to be somewhat standardized and the conclusion, as I understood, was that a mix of these approaches would provide the best solution. No problems were pointed out that might not be addressable by these techniques. In other words, the impression I got was that these techniques covered the entire problem/issue space related to 'file availability and consistency for mobile systems' and just a mix/combination of them would produce the optimal solution. Even if this were true, it didnt seem very convincing to me, as if so then this would mean that this topic would've become a dead issue which is not quite true. ------------------------------------------------------------------------------ My only problem with the paper is that you never define what `you' mean by `Moblie Systems'. Because this term can mean multiple things and because it was the centeral term in your summary paper I feel that defining this it was critical. Without this definition I was forced to spend the first few paragraphs forming this definition for myself. Over all I really liked your paper. I found it well structured, well written, and informative. I also wish you had included the techincal trends in your introduction. ------------------------------------------------------------------------------ The paper surveys an important topic in network systems with an emphasis on mobile systems. With the increasing use of networks and popularity of mobile systems, the problem of maintaining consistency and availability of files for a weakly connected or disconnected system is essential for continued computing. However, this problem was already dealt with at some length in papers cited in the refe- rence as early as 1989. The best part of the survey is that it gives a systematized and thorough discussion of the techniques. The evalu- ation of the techniques is adequate but the authors only give a rather general direction for improvements for the techniques. ------------------------------------------------------------------------------ In the paper, it justifies it is necessary to have better way to improve the File Availability and Consistency for Mobile connections (weakly/disconnect connection) However, I don't really see a why people want to do this, how often are people doing the work via mobile connections? (How often would a person download all the files he/she needs to do the job, and disconnect, and not return for a month or so?) I can see for user using modem to do the work (weakly connect) and suffer lots of problem due to the 'weak' connection, but what about people with 'disconnected connection?' It will be better if the paper actually talked about why people would want to do the work in this kind of environment and what are the technology trend now. Or what will this improve the performance/efficiency that will make users benefit and willing to start working in this computing model. It talks about lots of techniques to improve the File Availibity and Consistency, BUT, compare to what? There aren't any hard numbers as for Performance references, i.e. how much is this technicque improve the overhead 'compare' to other techniques. ------------------------------------------------------------------------------ This was a well-written paper that explored a number of techniques in a very readable fashion. Illustrations were useful. Authors might have spent a little more time justifying applicability of these techniques rather than relegating the arguement footnote. ------------------------------------------------------------------------------ Overall the paper was interesting to read and easy to follow. My only negative comment is that it is sometimes difficult to understand whether the presented ideas are original or simply restated from other works. ------------------------------------------------------------------------------ My review checks: 1-2 : poor. 3-4 : below average. 5-6 : good, above average. 7: excellent, how things should be. Review of "File Availability and Consistency for Mobile Systems" Import 1 - 7 5. While of some import, the increasing availability and low cost of network connections on the go (cell phones, RAM mobil net, 2-way pagers, etc) reduces the amount of time users spent weakly connected or disconnected, so this may of diminishing importance within the near future. Novelty 1 - 7 6. Lots of analysis, suggestions. Quality 1 - 7 5. Good content, but unpolished presentation, and some technical issues to fix. Overall 1 - 7 5. Good paper, with interesting suggestions for improvements on existing algorithms. However, couple things: 1) Who's Idea Is It, Anyway? I had a difficult time discerning what was the authors' opinions from what was taken from their sources. Would have liked if they said "We propose X" or something to that effect. 1.5) The suggestion on using server file hoardin to do replication. Not sure I understand why having a "very heterogeneous environment" would help. 2) A couple bald assumptions that may not be incorrect, but simply need more evidence. e.g. pg 3, column 2, paragraph 3: "ficus was later ported to disconnected operation, and *one must assume* this is how it is implemented" Also, a couple criticisms sounded almost irrational, e.g. pg 9, column 2, paragraph 5: "the conflict resolution mechanism in Ficus is poorly implemented" The reasons presented show that while the policy is not ideal, "poorly implemented" are too strong of words to descrbe the situation. One of several such statements that could be toned down. 3) Loose writing style, missing some "polish". "punted" is not an academic term. The paragraphs weren't indented, and the paragraph topics were not as descriptive as they should have been, e.g. "Primary and Secondary". Many sentences were wanting of clearer writing. It's and Its are not differentiated. While they may seem minor, they distract from the paper. Overall this paper could have done with another once-over to cleanup the little things. ------------------------------------------------------------------------------ Your comments on the paper. This is public comments that the authors of the papers will see. Provide feedback to improve their paper, etc. Good criticism for the various techniques used for replication, hoarding, and reintegration. The paper gave concise and logical observation to each technique and conclusions drawn from the observation are sound. The presentation of the paper didn't introduce much novel ideas. It didn't provide any new perspective to tackle the availability and consistency problem. Some of the terms used in the paper weren't clearly defined.