Apple Newton Internet Enabler Especificações descarregar pdf (Página 22)

June 1996 Newton Technology Journal

INTRODUCTION

Quality does makes a difference. It makes a difference to your customers and

ultimately to your bottom line. But quality assurance goes beyond just making

sure you find and fix bugs – that’s a given. However, you can’t test quality into

your application. Quality consists of meeting or exceeding customer

expectations in all facets of the user experience. Aspects such as the simplicity,

robustness, and utility of your application are critical to your success.

Whether you’re a large or small developer, a strong quality assurance

program can add a lot of value to your product. Besides giving you the

confidence to ship a well-tested product, Software Quality Assurance (SQA)

can act as the focal point for customer feedback upon which to build

increasingly customer-driven products.

Qualifying software on the Newton is not much different from doing so

on other platforms. There are three basic building blocks which comprise a

good Newton SQA program:

• an understanding of software development

• a foundation of general SQA practices

• knowing the specifics of testing on the Newton platform

SOFTWARE DEVELOPMENT

Your product development should be planned before you begin

implementation and you should execute to that plan. Marketing should

define the target customers and gather and relate their requirements.

Engineering should respond with a description of how those requirements

will be met by the software, both at the human interface and low level.

Based on the marketing and engineering requirements, SQA develops a test

plan defining how testing will be done to converge on a high quality

product. All of the teams must agree on what the end product will look like,

who it’s for, and what problems it solves. Never get so caught up in the

details of the project that you lose sight of the big picture.

Software development is an iterative process. What does it mean for

your software to reach the alpha, beta, final, and golden master milestones?

What metrics do you choose to measure your progress, and by what

milestone criteria do you judge the status of the project? These questions

are largely dependent upon the goals of the project. Choose metrics and

criteria with care. Minimalism is better than weighing your project down

with bureaucracy.

A well planned product development process provides the guidelines for

getting the work done, but software development is intrinsically dynamic.

Make sure your day-to-day decisions still reflect the project goals and

customer requirements.

GENERAL SQA PRACTICES

The following concepts underlie any good SQA program.

• Prioritized Test Coverage

Since you can’t test everything, prioritized testing is key.

What absolutely has to work from the customer’s perspective to make

your product successful? Concentrate more of your resources on those

features and less on others based on their relative importance.

Coverage should be partially driven from development engineering.

Ask the development engineers about the areas of their code and hacks

that concern them the most.

Look for interdependencies and assumptions. These might be bug-rich

areas.

Test any error-checking. Simulate the errors and validate the error-

handling.

Development and quality engineers should collaborate on creative ways

to break the code.

Weaknesses might be uncovered in code or design reviews.

Inexperienced programmers might require more test coverage, especially

in the system-specific features.

• Shared Knowledge

As much as possible, development and quality engineers should share a

common understanding of the features of the software and the underlying

code paths.

• Scientific Method to Isolate Bugs

The best quality and development engineers have internalized the formal

scientific method and subconsciously use it when debugging. “The real

purpose of the scientific method is to make sure Nature hasn’t misled you

into thinking you know something you don’t actually know.” [1] The

following is a terse description of the scientific method as applied to SQA:

1) statement of the problem

2) hypotheses as to the cause of the problem

3) test cases to test each hypothesis

4) predicted results of the test cases

5) observed results of the test cases

6) conclusions from the results of the test cases

• Managed Risk

Most decisions related to qualifying and shipping the product entail

varying degrees of risk. The challenge lies in correctly analyzing the risk and

taking steps to minimize it. Risk analysis should be undertaken seriously,

and any judgments should be based on facts, not conjecture. Always try to

think of creative ways to minimize the risk. Here are some examples:

Code changes

If there are code changes late into the project or near critical milestones,

there are some obvious ways to minimize the risk of the change:

Creating Quality Newton Applications

by Peter Murray, Apple Computer, Inc.

Product Development

1 2 ... 17 18 19 20 21 22 23 24

Comentários a estes Manuais

Sem comentários

Apple Newton Internet Enabler Especificações Página 22

Comentários a estes Manuais

Manuais e produtos relacionados com Software Apple Newton Internet Enabler