Hi, everyone, *Please read this mail carefully, it's important.* I have worked on the coding this 2 days, and now, I finish Part I code, After several tests, it run smoothly on my computer. Looks stable enough to do the Part II. The test result was attached to last weekly progress report. The Random testing strategy's average number of cases to find all faults is about *1000-1100*. I implemented the strategy proposed by George in Lecutre/Lab, which using 3 sets, Failure set, Successful set and Unused Set, to help select case. At beginning, all the cases are in the Unused set, and randomly pick one from it. If the case cause a failure, then put it into Failure set; if it's successful then put it into Successful set. From next round, we will look the Failure set first, if it's not empty, then randomly pick one to test until all of them are tested, then look at Unused Set, then Successful set. I tested my implementation, the average is about *800-900* cases. From the result, it's a little bit better than Random. I also implemented my strategy, which was proposed during 4th week's meeting, using length and number of different characters of both pattern and data to measure possibility of each case, I assume longer length and more different character will have more chance to cause an error. I finished it and tested it, it got very good performance, only *140 *cases to catch all the 21 faults. I think it's good enough for my marking about Part II. *If you have a better idea, then beat me using your result.* These are what I have done in this week. I will wrote a weekly progress report *TOMORROW*. *As said before, I will leave blank in section 2, Individual Contribution, for anyone who didn't send me this week's individual weekly report, that will means you did nothing in this week.* For your better marking, do something useful, and send me a reasonable result ASAP. If you didn't begin any work on project for this week yet. My suggestion is look at my code, and find a better strategy, at least beat George's strategy. Implement it by write your own XxxxxTestCaseProvider, test it, write a report for it, and send me both code's PATCH and report before I finished our group weekly report tomorrow. *3 Important things about our group project. * 1. I will change our *group meeting to Wednesday*, tell me your available time slot in Wednesday. Note, from 1pm - 8pm. Our meeting should be 1 hour, for some special case, it may be longer. Tell me your available time tomorrow. 2. Since I already finished the Software Testing Framework, it's quite easy to write your own strategy. For my SmartTestCaseProvider, it's only 74 lines(included blank line and comment line), for George's ThreeSetTestCaseProvider, it just 81 lines. For RandomTestCaseProvider, it's only 20 lines. My point is that it's easy enough for everyone write your own strategy by yourself, so write your own strategy as your contribution to this project. *From next week, everyone should do the project individually, unless some special job, we will have a compitition at the end of the week, the one get the best result will be the winner of the week. I will write the winner in the weekly report.* 3. During the meeting on next Monday, *7 April*, we will have following topic to do: 1) Confirm the *Wednesday meeting time.* 2) Make sure everyone can use *google group*, after confirmation, we will stop using the current mailing list. 3) Make sure everyone can use *google code hosting*, which my code is there. Everyone should be able to checkout, commit, or create a patch to me. (http://code.google.com/p/csci926project/) 4) To make sure everyone can write the code and debug, *Please take your laptop to the Monday's meeting.* 5) For those who are using Microsoft Windows system, please *DOWNLOAD *and *INSTALL *following software before you come the meeting, we will use this tools at Monday's meeting: a) TortoiseSVN (http://tortoisesvn.net/downloads) b) AnkhSVN ( http://ankhsvn.tigris.org/servlets/ProjectDocumentList?folderID=7315) c) MinGW ( http://sourceforge.net/project/showfiles.php?group_id=2435&package_id=240780 ) d) If you don't have Visual Studio 2005, then download and install Visuall C#/C++ 2005 Express Edition ( http://www.microsoft.com/express/2005/download/)[If you don't have bandwidth to download it, I can burn you a DVD, but let me know *BEFORE Monday*.] If there is anything I'm missing, write an email and let me know. Best Regards, Tao Wang -- Dancefire Microsoft Certified Technology Specialist CCNA http://www.dancefire.org/ I'm enjoy in Operating System, Network Security and Embedded System.