前几天写了篇博文“审一篇SCI稿件，整个人都不好了”，没有想到被编辑精选，而且引来了不少博友的关注，很多博友都提出了比较中肯的建议，如29楼的博友写了这样的评论：“对于一些低级错误，容易修改的不能作为拒搞的理由，审稿不能带个人情绪，文章应该较重创新点等”这些观点我是认同的。还有今天44楼的评论：“这也许是一个学生作者第一次投稿，应以鼓励为主，当然写成这样说明导师也没有好好看，或就没有给导师看就投了。可以指出他的错误，拒稿可以，但不能带有指责的情绪，更不能有居高临下的架势，谁都有从不会到会的过程，部分同意29楼的评论。”我也是认同的。

鉴于这篇文章有不少人关注，我感觉有必要再写一篇博文说明一下。这里主要强调一下博文与论文评审意见的不同。对于我来说，博文往往带有一定针对性或者倾向性，主要是为了吸引读者的眼球和关注度，当然这并不是我写博的初衷，但既然写出博文还是希望有更多的读者看到。如我的博文标题”审一篇SCI稿件，整个人都不好了”，其实审这篇论文，还是有点小激动的，毕竟是第一次审稿，怎么可能整个人会不好了呢？至于为什么要起这样的标题，正如前面所说的为了吸引眼球。

个人认为除非政论性的文章，不应该带感情倾向之外，其他叙述性的文章，还是带情感倾向比较吸引人，所以在接下的博文写作过程中，突出了自己的心情变化。

而论文评审意见是一件严肃的事情，不管作者的文章质量如何，最重要的一点就是尊重作者，以探讨学术问题为主，并且适当地给出一些个人建议，做到公正客观，不夹杂个人情感。

为了说明博文与实际的评审意见不同，现将个人写给作者的评审意见贴出（不知道这是否合适，如不合适请各位博文指出，我会及时将评审意见删掉）。

以下是我写给作者的评审意见： 

Reviewer Comments Directed to Author(s)

1. The paper presentation requires significant improvement. In its present form, it is difficult to read it. There are many typos, grammar and syntax errors that it is very difficult to enumerate them all. For example :

(1) namely the Relative Deadline, Worst Case Execution time etc.  -> namely the Relative Deadline, Worst Case Execution Time etc

(2) The Preemptions helps -> The preemptions help

(3) The Preemption threshold Scheduling (PTS) -> The preemption threshold scheduling

(4) The CPPTS algorithm provides...  and  The Cyclic Priority Pre- Emptive Threshold  scheduling (CPPTS), -> before you define the CPPTS, we don’t know what is CPPTS. You should denote it before using.

(5) Higher priority and lower priority -> higher priority and lower priority

(6) Each Task  is  assigned  -> Each task  is  assigned 

(7) if the  Interrupt latency goes -> if the interrupt latency goes

2. The notations or definitions is confusing or ambiguous. Such as: Let ‘’ be the list of task set to be served.  “Each  task   is denoted by “”.

3. The authors denote the the average number of preemptions and  the maximum number of preemptions , but they are not used in this paper.

4.  Most of the references are old. There are many papers on limited preempted scheduling, for example:

[1]Jinkyu Lee, Kang G. Shin, Preempt a Job or Not in EDF Scheduling of

Uniprocessor Systems. IEEE TRANSACTIONS ON COMPUTERS, 2014.

[2Esposito F, Yao G, Marinoni M, et al. Preemption Points Placement for Sporadic Task Sets[C]// 2012 24th Euromicro Conference on Real-Time SystemsIEEE, 2010:251-260.

[3Yao G, Buttazzo G, Bertogna M. Feasibility Analysis under Fixed Priority Scheduling with Fixed Preemption Points[C]// 2013 IEEE 19th International Conference on Embedded and Real-Time Computing Systems and ApplicationsIEEE, 2010:71-80.  

... etc.

5. The authors say “The  reduction  in  the  relative Deadline  reduces the number of preemptions thus reducing the interrupt latency.” But they don’t explain how the relative deadline reduces? And they don’t explain why the relative deadline reduces can cause in reducing the interrupt latency? This is very important problem in this paper.

6. How to assign the preemption threshold ? You method has advantage or not compared to the method in (M. Saksena and Y. Wang, ―Scalable real-time system design using preemption thresholds).

7. The proposed CPPTS algorithm uses the buffering time of the higher priority task to serve the tasks of lower priority in a cyclic manner. How to use the buffering time to serve the tasks of lower priority? before the high priority execution or after the high priority completion? Whether it will effect the feasibility of the algorithm?

8. Although the paper presents a new CPPTS algorithm, this algorithm is compared only to an EDF or RM scheduling algorithm, and not to other limited preemption algorithms. Therefore the comparison is clearly favorable to the algorithm presented. Please compare your algorithms and the experimental results to equivalent preemption algorithms. Such as:

  [2Esposito F, Yao G, Marinoni M, et al. Preemption Points Placement for Sporadic Task Sets[C]// 2012 24th Euromicro Conference on Real-Time SystemsIEEE, 2010:251-260.

  [4Marko  Bertogna  and  Sanjoy  ―Limited  Preemption  EDF  Scheduling  of  Sporadic  Task  Systems, IEEE  TRANSACTIONS  ON  INDUSTRIAL INFORMATICS, VOL. 6, NO. 4, NOVEMBER 2010

9. In table 1, TASKSET is ambiguous. The number stands for the number of task set? Or the number stands for the number of task in the task set? If the number stands for the number of task set, How many task in each task set?

10. Can you explain CPU utilization in table 1? Why not use the number of preemptions to evaluate the performance of the algorithm。