Test Design Guidelines for Reusability
Last Updated: JAN.10.2008
From: http://safsdev.sourceforge.net/sqabasic2000/TestDesignGuidelines.htm
- Purpose
- App Map and Test Table UTF-8 File Encodings
- App Map Chaining
- App Map Resolving Embedded Variables
- Designing Tests for NLS
- A Complete Dynamic Text Example Support NLS
Purpose
This document provides a set of guidelines that enable a tester to plan the development of test designs and related assets to maximize reuse. The intended context for this reuse is to maximize the potential for effective system\application configuration testing and localization testing, also known as NLS testing. Throughout this document we will refer primarily to NLS testing, but the concepts apply equally well to configuration testing and other forms of testing requiring reconfigurable text values.
The App Map features referenced here as App Map Chaining and App Map Resolving Embedded Variables are only available in versions of SAFS dated on or after Aug, 2006.
It is important to recognize that the single most important notion for making test designs reusable for NLS and configuration testing is that no localizable or configuration-specific text must exist in the test designs or test actions--the test tables. All such text must be managed in separate text resource files, which in the SAFS world means separate App Maps. The test designs and test actions will reference the text via variable references which will be retrieved from the App Maps opened for the target language or configuration.
App Map and Test Table UTF-8 File Encodings
(Need to respecify the following statement based on most recent experiences with Rational Robot and Rational Functional Tester evaluations.)
In order to properly represent mixed character data (some single-byte and some multi-byte) the UTF-8 character encoding format is supported by RRAFS and SAFS. All App Maps and Test Tables intended to be used for NLS testing should be encoded in the proper format required by the tools in use.
App Map Chaining
Note: App Map Chaining by default is not enabled in order to maintain backwards compatibility with existing tests. You must use the 'AppMapChaining' Driver Command to enable this feature.
C, AppMapChaining, ON
App Map Chaining is the process by which multiple App Maps can be opened and searched for referenced values. App Maps opened with chaining enabled are searched in Last-In-First-Out (LIFO) order. This allows the opening of scenario-specific maps--language or configuration, etc.--to be opened in addition to generic and shared maps. Maps opened later can have values that override or supercede maps opened earlier. In addition, shared maps can contain embedded variables (see below) that can be satisfied by the scenario-specific maps. For example, a shared map may reference an embedded variable containing localizable text that will be found in another open App Map--the one specific to the language being tested.
Example:
C, SetApplicationMap, FMStudio_NLS
C, SetApplicationMap, FMStudio_MAP
This design snippet shows two maps being opened. The first opened map (FMStudio_NLS) would contain the localized text specific to the language being tested. All entries in this map would be in the default [ApplicationConstants] section so that they could be retrieved as variables. The second map (FMStudio_MAP) would be the shared map used in all testing no matter what the language or configuration. The entries in this shared map would contain embedded variables whose values would be found in the localized map. Since the maps are read in LIFO order due to App Map Chaining, values will first be sought in the FMStudio_MAP. Embedded variables or other values not satisfied by that map will then be sought in the FMStudio_NLS map.
App Map Resolving Embedded Variables
Note: Resolving embedded variables in App Maps is not enabled by default in order to maintain backwards compatibility with existing tests. You must use the 'AppMapResolve' Driver Command to enable this feature.
C, AppMapResolve, ON
An App Map entry can contain references to variables whose values will be resolved to complete the entry. See Resolving DDVariables for Dynamic Recognition for more information on this capability. Essentially, the idea to keep in mind is that the shared maps will contain largely static text that does not need to be localized or changed for subsequent testing. Where some portion of an App Map entry does need to contain localized text that portion will be represented as an embedded variable and the ultimate value of that variable will be retrieved from a locale-specific App Map that is opened in addition to the shared App Map.
Example:
FMStudio_NLS_en (known as FMStudio_NLS):
----------------------------------------
[ApplicationConstants]
nlsHello="Hello!"
FMStudio_NLS_es (known as FMStudio_NLS):
----------------------------------------
[ApplicationConstants]
nlsHello="¡Hola!"
FMStudio_MAP:
-------------
HelloLink="Type=HTMLLink;HTMLText={^nlsHello}"
Now, let us assume that the test is run first in English and then later in Spanish. The first time the test is run the FMStudio_NLS_en (English) map is loaded along with the FMStudio_MAP. When the test goes to retrieve the value for the HelloLink component it will find the embedded ^nlsHello variable reference which must be retrieved. The ^nlsHello variable does not exist in the shared map so App Map Chaining will go look for the value in the English map. The value retrieved for ^nlsHello is "Hello!" and the resulting HelloLink value will be:
HelloLink="Type=HTMLLink;HTMLText=Hello!"
The second time the test is run the FMStudio_NLS_es Spanish map is loaded along with the FMStudio_MAP. This time the value retrieved for ^nlsHello is "¡Hola!" and the resulting HelloLink value will be:
HelloLink="Type=HTMLLink;HTMLText=¡Hola!"
As you can see, App Map Chaining with Embedded Variables is a powerful combination that enables us to deliver tests that can readily be migrated for NLS testing. However, to accomplish that it requires the Test Designers to design the tests and these assets with the goal of NLS reuse in mind.
Designing Tests for NLS and Other Forms of Reuse
Just as special design consideration must be given during software development to make an application support Internationalization, Localization, and Globalization, so too must our test designs receive this same level of design consideration. If the tests are not designed from the very beginning with this goal in mind then the process of going back and making the tests NLS-ready can be very difficult and time consuming.
The App Map processing features mentioned above allow us to use the same common practice used by software developers when developing software. That is, we must design our tests for NLS by separating textual data and other environment-dependent resources out of the tests themselves and into separate resource files. This means we want all environment-dependent text stored in environment-specific App Maps. Other environment-dependent resources like image files should be referenced through environment-specific entries in those same App Maps. In this way, the tests themselves do not change no matter which environment is being tested. When we go to test in a new environment we just make sure we use the appropriate App Map(s) and resources targetted for that environment.
It takes vigilance and attentive awareness to design a test that is truly NLS-ready. All environment-specific text and resources must be kept out of the design and moved into the resource files (App Maps). The designs will merely contain variable references that will ultimately be satisfied by lookups into the environment-specific App Maps. This not only includes language-specific text that might appear in object recognition strings, but also any language-specific string that might be evaluated or verified--such as an object property value, displayed text, or image benchmarks whose image may contain environment-specific text, legends, or formats.
An example of this might be a Label on a dialog box that informs the user that a certain process either succeeded or failed. The recognition string for the label may not require special handling as its Name or ID may be common across all environments as shown below:
StatusLabel="Type=Label;Name=lblStatus"
However, when we wish to verify the status displayed to the user in the Label we are going to be dealing with literal text that will likely be localized and needs to be stored separately in our NLS resource files or App Maps.
Problem:
T, StatusDialog, StatusLabel, VerifyProperty, "Text", "Success!"
Solution:
T, StatusDialog, StatusLabel, VerifyProperty, "Text", ^nlsSuccess
To make sure our tests are NLS-ready we are tasked to verify a few things:
- The Name of the Label does not change in other environments so it does not need to be localized.
- The name of the Text property does not change in other environments so it does not need to be localized.
- The displayed text in the Label will change in other environments so it does need to be localized.
Thus, our NLS-ready App Maps need to hold this localizable text so that the proper values are found at runtime:
FMStudio_NLS_en (known as FMStudio_NLS):
----------------------------------------
[ApplicationConstants]
nlsHello="Hello!"
nlsSuccess="Success!"
nlsFailed="Failed!"
FMStudio_NLS_es (known as FMStudio_NLS):
----------------------------------------
[ApplicationConstants]
nlsHello="¡Hola!"
nlsSuccess="¡Éxito!"
nlsFailed="¡Fallado!"
As you can see, it takes special attention and vigilance to carefully design our tests so that they are NLS-ready! The same level of care must be taken for all forms of displayed text, text to be input from the user that might be language-specific, and other resources containing environment-specific data like locale-specific currency, time, and date formats.
A Complete Dynamic Text Example Supporting NLS
- The Scenario:
- You wish to create a highly reusable step table (action) that will close any window in your application and then check the application state. We'll call this the "CloseAppWindow" table.
This can serve a number of purposes, including:
- High value test assets with reduced maintenance
- Supports dynamic features of the AUT
- Supports growing automated coverage of the AUT
- Existing automated tests are suitable for NLS testing
The Problem:For maximum reusability, CloseAppWindow must be able to recognize different windows according to their any-language Caption, so no part of CloseAppWindow, or any of the tables that call it, can contain language-specific text.
The Solution:Use App Map Chaining, App Map Embedded Variables, and Application Constants.
1. A few examples calling CloseAppWindow with the Caption parameter using Application Constants:
T CloseAppWindow ^Caption=^ReportsCaption T CloseAppWindow ^Caption=^HelpCaption
T CloseAppWindow ^Caption=^AboutCaptionThere is no localizable text calling CloseAppWindow
- .
2. The implementation of the CloseAppWindow step table:
T AnyWindow AnyWindow CloseWindow
...followed by important status and error checks There is no localizable text in CloseAppWindow.
3. The Primary App Map using App Map Embedded Variables:
MyApp.map:
----------
...
[AnyWindow]
AnyWindow="Type=Window;Caption={^Caption}"
... There is no localizable text in the primary App Map.
4. The NLS App Maps for App Map Chaining for Application Constants:
4a. MyApp_NLS_en (known as MyApp_NLS):
--------------------------------------
[ApplicationConstants]
ReportsCaption="Reports"
HelpCaption="Help"
AboutCaption="About" All localizable text is confined to the NLS App Map. 4b. MyApp_NLS_es (known as MyApp_NLS):
--------------------------------------
[ApplicationConstants]
ReportsCaption="Informes"
HelpCaption="Ayuda"
AboutCaption="Sobre"
This sample design shows maximum reusability. The same test used to automatically test the English translation of the application can be used to test all other localizations. And the amount of localization necessary for each translation has been minimized to a single NLS App Map file.
Test Design Guidelines for Reusability的更多相关文章
- Database Design Guidelines
Database Design Guidelines Principles Support popular databases Name Style Table Name Style: Pascal ...
- [书籍]重温《Framework Design Guidelines》
1. 前言 最近重温了<Framework Design Guidelines>. <Framework Design Guidelines>中文名称为<.NET设计规范 ...
- Dynamic Library Design Guidelines
https://developer.apple.com/library/mac/documentation/DeveloperTools/Conceptual/DynamicLibraries/100 ...
- SAP成都C4C小李探花:浅谈Fiori Design Guidelines
Jerry: 我和周帅认识不久,自去年7月SAP成都研究院Cloud for Customer(以下简称为C4C)开发团队组建至今,根据这段时间和周帅愉快的合作经历,我觉得如果把周帅比作我读过的小说里 ...
- Three Sources of a Solid Object-Oriented Design
pingback :http://java.sys-con.com/node/84633?page=0,1 Object-oriented design is like an alloy consis ...
- Android Programming: Pushing the Limits -- Chapter 4: Android User Experience and Interface Design
User Stories Android UI Design 附加资源 User Stories: @.通过写故事来设计应用. @.每个故事只关注一件事. @.不同的故事可能使用相同的组件,因此尽早地 ...
- The design of a distributed variant of Plato framework to support collaborated editing
A first thought system architecture (pulling mode) is one that the server doesn't keep client inform ...
- 【android design】android常用设计资源
一.概述 大部分程序员擅长开发,但是对于设计却知之甚少.这直接导致,程序员在初期开发出来的应用(大多为兴趣或实用导向)中看不中用.因此,有必要搜集整合一些设计资源,这样既能减轻程序员在设计上所耗费的时 ...
- API Design
REST API Design Guidelines V 1.0.201208 Draft 5 Last Updated: 08/31/2012 1 简介 本文档旨在规范REST API的 ...
随机推荐
- 黑马程序员_Java基础:集合总结
------- android培训.java培训.期待与您交流! ---------- 一.集合概念 相信大家都知道,java是一门面向对象的编程语言,而对事物的体现都是以对象的形式,所以为了方便对多 ...
- ArcGIS线转面的方法
ArcGIS作为GIS软件中的龙头老大,其功能是非常强大的,但是如果作为一个初学者,其部分常用的重要功能不容易掌握,今天就讲一讲在矢量化时非常重要的功能:线转面. ArcGIS在进行大范围的矢量化时一 ...
- PYTHON学习之路_PYTHON基础(8)
学习内容: Python模块介绍 1.经典类 or 新式类 2.抽象接口 3.静态方法.类方法.属性方法 4.反射 5.异常处理 6.socket编程初识 7.用socket实现get.put文件等功 ...
- 人脸识别经典算法三:Fisherface(LDA)
Fisherface是由Ronald Fisher发明的,想必这就是Fisherface名字由来.Fisherface所基于的LDA(Linear Discriminant Analysis,线性判别 ...
- iOS开发流程总结
本文由社区会员umyueyue分享 以下是会员umyueue总结的iOS开发流程以及学习中的资料分享. 流程:注册.开发.真机测试.发布以及上线. iPhone iOS 4从注册到app上线开发流程 ...
- C2第十次解题报告
看过题解后如果觉得还算有用,请帮忙加点我所在团队博客访问量 http://www.cnblogs.com/newbe/ http://www.cnblogs.com/ne走迷宫wbe/p/406983 ...
- EI Index
Journal of Software (JoS). China. [link]
- iOS Wi-Fi
查漏补缺集是自己曾经做过相关的功能,但是重做相关功能或者重新看到相关功能的实现,感觉理解上更深刻.这一类的文章集中记录在查漏补缺集. iOS 开发中难免会遇到很多与网络方面的判断,这里做个汇总,大多可 ...
- C++混合编程之idlcpp教程Lua篇(7)
上一篇在这 C++混合编程之idlcpp教程Lua篇(6) 第一篇在这 C++混合编程之idlcpp教程(一) 与LuaTutorial4工程相似,工程LuaTutorial5中,同样加入了四个文件: ...
- the diff typeof and instanceof
instanceof和typeof都能用来判断一个变量是否为空或是什么类型的变量. typeof用以获取一个变量的类型,typeof一般只能返回如下几个结果:number,boolean,string ...