做事要像老牛一样踏实,思维要像猴哥一样灵活,生活要像熊猫一样舒服。

发布新日志

  • Validated System

    2009-03-11 14:05:05

    No 1: Difference between Validated and NonValidated:

    (1)   Validated System------Closed System

    (2)   Non Validated------Open System

     

    No 2: Which rulers does validated system contain?

    (1)   Every user has to be unique

    (2)   Every signature should be unique identified

    (3)   Every record has two signature (autor and reviewer)

     

    No 3: How to manage validated system?

    (1)   Administratered

    (2)   Audit Trail

    (3)   Person are trained

    (4)   Individuals are responsible

    (5)   Data cann’t be manipulated

    (6)   Reproducible

     

    No 4: Which method does Validated system use?

    (1)   Include screen shot in every resulting step

    (2)   Convert finalized test case to PDF

    (3)   Print out

    (4)   Which way execute Test Case

     

    No 5: How to create the protection for PDF?

    (1)   Install Security Add-in

    (2)   File--->Document--->Security---->Certificate Security---->Save these settings as policy

    (3)   After creatig protection, you can click “sign”---->”autor”/”reviewer”, and then you can change or write review comments

     

    No 6: Where identifies “Validated/Non Validated”?

    (1)   21 CFR (part 11), CFR is the abbreviation of “Code of Federal Regulation”, it is created in 1999 by Cliton

    (2)   FDA------Food & Drug Administrator

    (3)   21 CFR (part 11) is about electronic signature & Elec Records

  • Requirements Analysis & Prototyping Tools

    2009-03-11 09:38:50

    No 1: What does BA do initially?

    (1)   Interview users

    (2)   Gather requirements--->mainly on functional requirements

    (3)   Design specification

    (4)   Functional specification, like user case

    (5)   SME--->Subject Matter Experts

     

    No 2: What does BA do in next interview?

    (1)   Interview

    (2)   Defects & Enhancement

    (3)   Functional Enhancement & Priorities

    (4)   UATS

    (5)   Training

     

    No 3: What is BA’s skill?

    (1)   Communication (Document)

    (2)   Domain knowledge

    (3)   Training

    (4)   Analysis & Summary

    (5)   Has to be a good listener

     

    No 4: What is BA responsibility for?

    (1)   Document

    (2)   Consistant communication & clarification

    (3)   Bridge between customer & development

    (4)   Conduct JAD(Junit application Design) session

    (5)   Gather requirements

    (6)   User Case

    (7)   Prototypes

     

    No 5: Which technology skiils does BA master?

    (1)   MS word (Document ability)

    (2)   UML (Drawing tool)

    (3)   Excel

    (4)   Prototype tools (Axure RP Pro v4.4.1.745, EnterpriseArchitectv7.0.816, ER studio 7.0.1 and so on)

     

    No 6: What is prototype?

    (1)   What I hear, I forget

    (2)   What I see, I remember

    (3)   What I do, I understand

    -------Lao Tse

    No 7: Which kinds of tools for designing user case?

    (1)   Open source: StartUML, Umbello--->Linux, AndroMDA, BOUML, ArgoUML

    (2)   Licensed Tools: MS Visio, Rational Rose, Visual Paradign, SmartDraw, etc...

     

    No 8: UML must contain three components:

    (1)   Actor

    (2)   Include/extend

    (3)   Use case

  • Open-Source Testing Tool

    2009-03-10 17:25:38

    No 1: We can look for Open-Source Tool for testing from the following web site:

    (1)   www.opensourcetesting.org

     

    No 2: Open-Source testing tool:

    (1)   Selenium, it is a functional testing tool, and suit for agile testing. Its’ advantage: IDE, Core, Remote Control; Its’ disadvantage: limited support, it base on FireFox Web browse

    (2)   QATraq, testing management tool

  • Performance Testing

    2009-03-10 17:15:12

    No 1: Before we start to do performance testing, we need do the following things:

    (1)   Plan

    (2)   Identify trouble points (Bottlenecks)

    (3)   Identify the hardware

    (4)   Identify the technology

     

    No 2: If we do performance testing for web page, we should clean cookies before testing.

     

    No 3: There are two kinds of services:

    (1)   Web service: Apache

    (2)   Application service: WebLogic, WebSphere

     

    No 4: Performance testing, usually, we check the following points:

    (1)   Hits/second

    (2)   Throughput: how much pages be sent to service

    (3)   Response Time: t1+t2+t3+t4; If there isn’t web service,then it is equal to t5+t6

    (4)   Number of register user

    (5)   Number of transaction record

     

    No 5: How to use WebLoad to do performance testing? You can refer to the following steps:

    (1)   Creating and editing scripts

    (2)   Creating & running Load test

    (3)   Analyzing load test results

     

    No 6: In result report, the results is passed, however, there is problem, why we say like that?

    (1)    “Passed” status means that the action has finished, however, there is issue in performance.

    (2)    Choose “goal-oriental” model, and then create checkpoint

     

  • Testing Instance

    2009-03-09 17:49:54

    No 1: Create test sets for one application system, you can do as the following:

    (1)   Regard one page as one application screen

    (2)   Regard one section of one page as child group

    (3)   Create test case in each group

     

    No 2: When we test Financial Website, we must consider the following three points:

    (1)   Testing session, for example: clicking “back” key

    (2)   Checking whether can add user information page into favorite, for example: login financial system, and then add some page into favorite. If can be added, it is a bug.

    (3)   Testing security

  • How to use QC to manage testing

    2009-03-09 17:16:33

    No 1: As a QA leader, if you need QC to manage testing work, you can do as the following:

    (1)   Creating project in QC Site Administration

    (2)   Setting up groups in QC Tool Customization

    (3)   Adding users to group

    (4)   Adding fields in QC

    (5)   Importing Requirements

    (6)   Maintaining clarification logo

     

    No 2: QTP object repository usually likes the following:

    (1)   There are two kinds of object repository: local and share

    (2)   Object repository framework: common object repository, common script, common parameter

    (3)   Object repository Library: Pro object repository, Pro test script, Pro test parameter

     

    No 3: If you have built connection between QTP and QC, however, QTP script. cann’t run in QC, you can do as the following:
    (1) In QTP, clicking “Tools”--->”Options”--->”run”

    (2) In “run” page, checking on “Allow other Mercury products to tests and components”

    (3) In QC, run QTP scripts again

     

    No 4: Create project template in QC Site Administration

    (1) In browse address, input “http://XXX.XXX.XXX.XXX:8080/qcbin/”

    (2) Clicking “Site Administration”, and open “Site Administration” page

    (3) Selecting one domain

    (4) Clicking “Create Project”, then you can set fields and save it as project template

     

    No 5: Identify testing process by QC; you can identify some fields if you need.

    (1)   Requirement, before import requirements, we need to add these fields:

    1’ out of scope---value (review status)

    2’ application screen---field (parent/group name, filter nonrequirements)

    3’ change priority---value, opposite to defect’s priority (1-urgent....5-low)

    4’ review comment

    5’ Release

    (2)   Test Plan, before making test plan, we need to add these fields:

    1’ peer review

    2’ Developer review

    3’ sign off

    4’ assigned

    5’ reviewer

    6’ comments

    7’ version

    8’ automatic---Y/N

    9’ Application Screen

    10’ priority

    (3)   Test lab, before executing test case, we need to add these fields:

    1’ environment---field, it contains these values: test, stage, development, production

    2’ Instance

    (4)   Defect, if we need, we can add these fields:

    1’ reviewer

    2’ enhancement---Y/N

    3’ assigned by

    4’ group, it contains: Developer, tester, user

     

    No 6: Who manage requirements?

    (1)   BA manage requirements

    (2)   QA can review it

    (3)   QA cann’t remove requirement

     

    No 7: How to distribute permission to each group?

    (1)   tools--->customization--->set up groups

    (2)   choose one group, click “view” button

    (3)   click “defect”, select “status”, you can change the transition rules of status

     

    No 8: How to create test lab?

    (1)   Create test set

    (2)   Click “Select Tests” and select “test case”

    (3)   Click “Add tests to test set”

     

    No 9: How to create test plan tree?

    (1)   Usually, we set up these groups in test plan: release, regression, automated, removed

    (2)   We use control no(010...) to identify different level

    (3)   We can do as the following figure:

     
  • Project Customization in QC

    2009-03-06 14:00:58

    No 1: If you need special fields for requirements which don’t exist in system default fields, you can do as the following:

    (1)   Clicking “Tools”--->”Customize”

    (2)   Clicking “Customize Project Entities”

    (3)   Opening “Requirement”--->”User Fields” folder

    (4)   Clicking “New field” button at the bottom of the page

    (5)   In the “New field” page, inputing “Field Lable” what you want

    (6)   If select “Field Type” which is “lookup list” type, then it will display “Lookup List” page. And then you can do the following:

    1’ If it has existed the list what you want in the drop-down list, you can select it and click “save” button;

    2’ If it has existed the list name what you want in the drop-down list, however, you need to change the fields of the list, you can click “goto list” and click “New Item”, then save it;

    3’ If there isn’t the list what you want, you can click “New List” button.

     

    No 2: If you need to change “Project List”, how do you do? You can do as the following:

    (1)   Clicking “Tools”--->”Customize”

    (2)   Clicking “Customize Project List”

    (3)   Selecting the list name in the “Lists” drop-down list

    (4)   If you need, you can click “New Item” button to add some items

    (5)   If you need a new list, you can click “New List”

    Usually, these lists can be used in every component (Requirements, Business Components, Test plan, Test lab, Defects)

  • Severity VS Priority

    2009-02-27 10:27:21

    Question: If there are many defects, should we fix high priority defect or severity defect?

    In my opinion, we should fix the higher severity defects. Because this kind of defects would affect system running and testing schedule, the other reason is that we identify the priority according to severity. The defect is more serious, its’ priority level is higher.

    Pathik’s answer: It maybe higher severity or higher priority, that depends on person. If he is the customer, he should like to fix the higher priority defect. Because he think the higher priority feature is important. If he is the developer, he need to fix the higher severity defect, because it will affect the system’s running and testing. That is to say, fixing which kind of defect first, it depends on situation.

    Usually, customer identifies the priority like that: the feature is more important, the priority is higher. However, in development team, we used to identify the priority like this: the severity is higher, the priority is higher, and we would fix defect according to the priority.

  • 傻瓜使用JProfiler监视JSP实例详述

    2009-02-10 13:52:06

    一,环境搭建

    1.       安装jdk-6u11-windows-i586-p.exe,安装在D:\JavaServer

    2.       安装apache-tomcat-6.0.18.exe,安装在D:\JavaServer

    3.       安装MyEclipse_6.5.1GA_E3.3.2_Installer.exe

     

    二,配置MyEclipse中的Tomcat

    1.       打开MyEclipse

    2.       Windows——〉Preferences——〉MyEclipse Enterprise WorkBench——〉Servers——〉Tomcat——〉Tomcat6.x(根据你所安装的版本而定)

    3.       Enable Tomcat Server,并且选择Tomcat的安装路径,如图1

    1

    三,创建测试项目

    1.       新建Web项目mytest:

    (1)        打开MyEclipse

    (2)        New——〉Project——〉MyEclipse——〉Web Project(Optional Maven support)——〉单击“Next,显示图2,在图2中输入project mytest,选择JSTL1.1支持。

    2

    2.       创建类

    1mytest下的src右击新建一个包,包名cn.test

    2选中cn.test,右击新建两个类TestMain.java TestBean.java

    package cn.test;

    public class TestBean {

      String name = "";

    }

     

    package cn.test;

    import java.util.ArrayList;

    public class TestMain {

    public static ArrayList list = new ArrayList(); //存放对象的容器

    public static int counter = 0;    //作统计用

        }

    3.       创建Jsp页面

    (1)       mytest工程中,选中WebRoot文件夹,右键菜单New->JSPFile Name修改为init1.jsp,其他默认,点击Finish按钮创建。编辑init1.jsp文件内容:

    <%@ page language="java" import="cn.test.*" pageEncoding="ISO-8859-1"%>

    <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">

    <html>

      <head>

        <title>init</title>

      </head>

      <body><%

       for(int i=0;i<10000;i++){

           TestBean b = new TestBean();

           TestMain.list.add(b);      

       }

       %>

       SIZE:<%=TestMain.list.size()%><br/>

       counter:<%=TestMain.counter++%>

      </body>

    </html>

    (2)       init2.jspinit1.jsp一模一样即可(后面有用)。

    4.       部属Web应用程序

    这里就简单点了,我们选中mytest工程的根目录,右键菜单MyEclipse->Add and Remove Projects Deployments…,在打开的对话框中确保ProjectsWebTest,点击Add按钮,在弹出对话框的Server中选择Tomcat 6,下面的两个单选项,Exploaded ArchivePackaged Archive分别是目录方式部属和包方式部属,这个都是由MyEclipse来做的,我们使用目录部署方式,不用打成war包,其他的不管它了。点击Finished按钮回到上一个页面。每次添加修改jsp,先删除原先的,再添加最新的。

    5.       运行调试程序

    Eclipse的图形菜单栏找到 (这个按钮左边那个是重新部署按钮,右边是自掉的浏览器),点击下拉按钮,选择Tomcat 6,然后点击Start菜单,启动Tomcat

    Console窗口中提示Tomcat启动成功,我们就可以运行我们的程序了。打开一个IE浏览器窗口,输入http://localhost:8080/mytest/init1.jsp,我们可以看到如下信息:说明运行成功

     

    SIZE:10000
    counter:8

     

    三,启动JProfiler

    参照http://blog.csdn.net/djy1135/archive/2008/04/18/2304465.aspx 进行配置使用。注意,启动start_tomcat_jprofiler.bat之前,最好把myeclipse中的tomcat服务停了。

     

    四,参照

    MyEclipse开发调试WEB应用实例入门 - 启天的技术小屋[CSDN Junnef]...

    http://www.360doc.com/content/070209/01/11553_360989.html

    JProfiler学习笔记

    http://blog.csdn.net/djy1135/archive/2008/04/18/2304465.aspx

  • JProfiler New ServerIntegration出现的问题

    2009-02-09 18:05:55

    参照http://blog.csdn.net/djy1135/archive/2008/04/18/2304465.aspx进行配置New ServerIntegration。出现了以下问题:

    (一)我本机使用的Tomcat6.0jdk1.6.0_11。配置完,启动Apache Tomcat6.x(with Tomcat6.exe) on localhost(4),产生一些警告,并disconnected JProfiler.

    原因:MyEclipse中启动了Tomcat,外部也启动了Tomcat,端口号发生冲突

    解决方法:打开目录:D:\JavaServer\Tomcat6.0\conf下的server.xml文件,修改下列端口号:

    1<Server port="8105" shutdown="SHUTDOWN">

    2<Connector port="8180" protocol="HTTP/1.1"

                   connectionTimeout="20000"

                   redirectPort="8443" />

    3<Connector port="8009" protocol="AJP/1.3" redirectPort="8443" />

    将原来的8080”修改成以上端口号,保存修改后重新启动myeclipse中的Tomcat,显示“信息: Server startup in 1226 ms”表明Tomcat启动成功。

     

    总结:New ServerIntegration的时候须注意:(1启动myeclipse中的Tomcat;(2)不要在外部启动Tomcat;(3)保证端口号的唯一。

     

    (二)IE地址栏中输入:http://localhost:8080/test/init1.jsp,提示“HTTP Status 404,网页没打开。

    原因:网址输错了

    解决方法:在IE地址栏中输入:http://localhost:8080/test/Init1.jsp

    显示下图:

     

    执行三次,我们可以在内存视图中看到cn.test.TestBean对象被创建了30000

  • 用Tomcat5.5.27配置JSP运行环境

    2009-02-04 10:23:46

    配置参考网址:http://blog.chinaunix.net/u/306/showart_76276.html

    出现了以下问题:

    (一)  以下两个网页打不开:

    http://localhost:8080/admin/index.jsp  http://localhost:8080/manager/html

    原因:因为tomcat5.5在默认的包中是不包含Administration Web Application的,所以在地址栏输入“http://localhost:8080/admin”将看不到管理界面效果。对于tomcat5.5来说,需要下载Administration Web Application这个包,下载的地址还是http://tomcat.apache.org/download-55.cgi

    解决方法:下载apache-tomcat-5.5.27-admin.zip后,解压,apache-tomcat-5.5.27-admin\apache-tomcat-5.5.27目录下的文件直接覆盖到该目录下C:\Program Files\Apache Software Foundation\Tomcat 5.5。最后只要在\apache-tomcat-5.5.26\conf下将tomcat-users.xml中加入 <user username="admin" password="admin" roles="admin,manager"/>即可,其中用户名和密码均为“admin” 重新启动Tomcat以后,你就可以使用该用户来登陆如上的两个管理工具,通过Web方式进行Tomcat的配置和管理了。

    (二)  Tomcat服务启动后不久自动停止

    原因:可能是JDK(jdk-1_5_0_08-windows-i586-p)Tomcat(apache-tomcat-5.5.27)的版本不兼容

    解决方法:卸载JDKTomcat,重新安装最新版本。先安装jdk-6u11-windows-i586-p.exe,再安装apache-tomcat-6.0.18.exe。安装过程中注意以下几点:(1)为了方便JDKTomcat安装在D盘根目录;(2Tomcat安装过程中设置admin用户的密码”admin”;(3JDK5.0Tomcat5.0以后的版本不需要设置环境变量,如果设置环境变量也可以。输入:http://localhost:8080,启动成功。
  • 总结2008年,展望2009年

    2009-02-03 10:42:46

    2008年对我来说是打基础的一年,取得了一些进步,但并不算大的进步。蔡学銿把IT学习能力分成五等:幼稚园小学中学大学研究所。对号入座,目前我应该处于“小学”这一等级。

      我觉得做项目是很能锻炼人,对于各方面技术水平的提高也很有帮助。2008年主要参加了两个项目:乐器租赁网站,营销策略项目。

      首先,总结一下在乐器租赁网站项目(200801——200804)中的收获。该系统测试主要由我负责。从了解需求,制定测试计划,设计测试用例,执行系统测试以及回归测试。我觉得这个系统测试做得很不好,在这次测试中发现了好多测试技巧方面的问题。

    (1)       测试时间估算不准确。由于Web测试经验欠缺,测试时间估算很不准确。该系统分前后台,前台包括下订单检索用户管理等功能,后台包括产品管理,优惠券管理,订单管理,用户管理,系统参数管理等。由于前期需求花费时间较长,开发结束就剩两周的测试执行时间,一个人测试人员需要执行系统测试性能测试,时间很不足。感觉做得手忙脚乱的。我觉得为了做好以后的测试估算应该做到一下两点:第一,使用文档准确统计完成每项测试任务的时间,这些数据有助于测试的估算;第二,学会分析利用客户需求,初步制定测试需求。这次测试,让我深深体会到测试估算不准确带来的风险。

    (2)       测试用例形同虚设,没有真正的使用。这个问题的存在主要是以下原因:不够主动;无人监督测试用例使用情况。不好好使用测试用例的后果就是,可能会存在测试不全面的风险。我觉得对于这种情况可以采取以下措施:项目组其他人员评审测试用例;测试经理监督测试用例执行情况。

    (3)       性能测试做得很低级。该系统的性能测试主要测试了前台多人提交订单,以及检索的性能。性能测试首先需要对客户提供的性能需求进行分析挖掘,做这个工作需要具备一些性能测试方面的知识以及行业背景的了解;选择合适的性能测试工具以及测试方法;对系统开发语言以及环境的了解。通过这次实践,让我感到性能测试知识的欠缺。

    (4)       测试任务执行优先级。时间比较短,需要测试点很多,如何安排好测试的先后顺序可以很好的提高测试效率。这个问题在回归测试的时候遇到了。我觉的应该先测试系统核心功能,然后再好好测试一些非主流的功能。但是,我当时是先测试比较闲的开发人员的bug,然后再测试比较忙得开发的bug。最后导致核心功能花费时间不够。大大影响了测试的质量。

      虽然,这次测试存在好多问题,但是收获还是蛮多。比如,尝试了在项目中使用QTP自动化辅助测试,对于自动化测试的认识有了大幅度的提高,为QTP更深入的学习做了很好的铺垫;合理的使用了VSTS进行缺陷的管理;与开发沟通比较及时到位等。

     

    乐器租赁网站项目结束后,参加了营销策略项目(200805——200810),这期间主要做了以下事情:

    (1)       搭建 工作环境, 熟悉需求

    (2)       开发设计测试用例,每天做基本功能的测试。另外,及时测试开发新增的功能以及回归测试

    (3)       开发测试报告记录测试细节,开发buglist.xls跟踪测试过程中发现的bug

    (4)       使用大批量数据运行算法,进行压力测试,记录服务器性能,检查log及时跟开发汇报bug

    (5)       使用客户提供的样品数据运行算法并坚持导出/分析运行结果,及时向开发汇报问题.

    (6)       参加database同步测试,检查DB log

    (7)       整理测试用例,算法的流程,协助开发完成算法流程图 

    在此期间,测试技巧,测试用例开发,英文写作水平得到了一定提高,并且学会了编写简单的mysql存储过程,性能测试理解得到进一步加深。同时也发现了一些不足,与开发的沟通不够主动,性能测试知识还需进一步学习。

     

    退出营销策略项目后,主要进行了一些学习。大致学习了以下内容:

    (1)       测试方法论。经常访问51testing,SQAforum网站,及时了解软件测试的最新动态

    (2)       深入学习QTP的知识,使用简单的自动化测试框架设计自动化测试用例,练习使用常用的VBscrīpt函数

    (3)       初步了解java性能测试工具JProfiler

    (4)       坚持英语新概念三的学习

    (5)       积极参加组内培训,并组织了一次 QTP入门实例培训

  • 关于如何更高效地进行回归测试的一些想法

    2008-12-08 16:43:59

    看到这个问题,好多人觉得这是测试人员的事,其实我觉得如何更高效的进行回归测试应该是开发人员和测试人员共同的事。

    为什么这么认为呢?举个简单的例子:一些开发人员修改完bug,他们觉得自己的bug都已经fixed,然后兴高采烈地通知配置管理员发布新的版本,结果测试人员刚跑几个模块,系统就崩溃了。这怎么让回归测试进行下去呀?所以我觉得要进行高效的回归测试,按角色分

    应该从以下几个方面着手。

     

    首先,开发人员应该做到以下两个方面:

    第一,   开发人员在发布新的版本之前要做smoke testing,尽可能早的发现一些影响测试的严重bug;

    第二,   开发人员在修改bug状态的时候,要注明修改了哪个模块的哪些函数,这些信息有助于懂代码的测试人员去分析判断该bug是否真的修复好并对系统产生哪些影响。

     

    其次,测试人员应该尽可能的从以下几个方面着手:

    第一,   要熟悉系统的业务流程。对于该bug(或新增功能)的业务需求以及关联模块要很清楚,可以尽快进入测试状态并保证测试的质量

    第二,   及时更新测试用例,保证执行的测试用例是最新的

    第三,   要掌握测试用例的优先级别,也就是分清孰轻孰重,把时间花在刀刃上。对于优先级高的功能优先并充分测试,时间允许的前提下再测试优先级低的功能

    第四,   借助自动化工具测试相对稳定的功能

    第五,   沟通,及时与开发人员进行有效的沟通,更多地了解业务及系统,及时反馈测试情况

    第六,   有效的测试管理,作为测试经理应该对于自己的组员有足够的了解,根据测试人员的技能,合理分配测试任务

    最后,我觉得测试人员应该熟悉系统开发的语言。

     

    以上是本人的一些想法,希望大家补充啊!
Open Toolbar