Software Development Guide

By Sandeep Desai (http://www.thedesai.net/)

Quality of code is inversely proportional to the amount of code duplication

Successful projects

• Proper vision
• Sponsorship
• Agile Development
• Well defined boundaries
• Flexible developers, sometimes requirements are not very clear
• Developers and Managers, who are interested in the project
• Automated builds and tests daily
• QA considered as important as development
• Opportunities for developers to learn new APIs
• Plan for upgrades
• Flexible Design based on Patterns
• Schedule time for Refactoring
• Use Code Analyzers for code coverage, dependency analysis

Typical problems in projects that fail

• Project Plan does not have proper vision or timeline
• Goals are too optimistic
• Project is too large
• Poor team morale
• Team roles are not clear
• No sponsors for project
• Not enough methodology
• Too much methodology
• Inexperience management
• Inadequate technical skills
• Too many meetings

Coding style

• Don’t write unnecessary comments
• Follow the www.nejug.org Coding Guidelines for Java
• Refactor code, since you will rarely get it right the first time

Teamwork

· At least two people know all the code, so that there is no single point of failure

· Have proper code reviews for major code submission transactions

Architect

• Knows and evaluates APIs, frameworks, 3^rd party services
• Know Open source resources available
• Recommends APIs, frameworks etc for a project
• Knows list of reusable components and services within the organization
• Helps in identifying high risk parts of the project
• Tracks standards
• Creates knowledge base of projects
• Feedback from finished projects

Individual

• Help everybody
• Answer all questions, even if they seem stupid.
• Take responsibility for your mistakes

• Enthusiastic and fun to work with.
• Eager to learn and implement new technologies
• Good typing speed
• Should have some certifications such as

Manager/Tech Lead

• There should be a project lead who understands the code and can maintain code quality
• Remove any impediments that impact developer productivity
• Understand which bugs are important

Organizations

· Allocate time for learning new skills

· Hire fewer better programmers and pay them well. A good programmer can be 10 times more than a mediocre programmer

·

Questions to ask when taking over a project

Initial issues

· Architecture

· Project Documentation (Design documents

· Code

· Database schema and code scripts

· Database Data

· Other Data such as XML files

· Install and Deployment scripts

·

Taking Over Project

• Code Documentation
• Code Size
• Code Quality
• Code Modularity
• Build Scripts
• Test scripts

Joining new Organization

• Organization Philosophy
• Development Philosophy (Is it Agile, Is usage of new technology encouraged, Bureaucracy)
• Office politics
• Location (in HQ or satellite office)
• Priority of project within the organization
• Growth Rate
• Profitability
• Upper management

References

Thinking in Java 3^rd Edition

Programming Languages

The three types of programming languages in popular use are Compiled General Purpose such as C, C++ and Java. High level task specific languages such as SQL, perl, HTML Perl, SQL and data definition languages such as XML.

The more esoteric and less common type of languages are functional languages like Lisp. Machine Code (Assembly language) for the Processor typically useful for writing device drivers.

A language is either compiled to machine code that can be directly understood by the processor or into some form of intermediate code which is then either interpreted or translated to machine code while the program is running. A language can also be interpreted such perl, SQL where the program is run directly.

Programming languages such as SQL can also be embedded into another language such as C, C++ or Java.

Types of Languages

• Programming like Java, C
• 4GL like SQL
• Structured Documents like XML

The Basic Language features are

• variables & scalar datatypes
• scoping (global, static, local)
• arrays
• new datatypes (structs)
• control (if/for/while)
• functions (pass by value, pass by reference, return values)
• Memory management (Heaps, stacks)
• Program call stack
• error handling
• namespace
• APIs
• String
• Persistence (Files and Databases)
• Process Management (Process/Thread and Inter Process communication)
• Networking

· OOP

• class/object
• inheritance
• polymorphism
• exception handling
• RTTI
• Aspects (Cross Cutting Concerns

Language Tools

• compiler (direct machine code or platform independent machine code)
• Linker
• Run

Data Structures and Algorithms

• Stack
• Queue
• Hashtable/HashMap
• Linked List
• Tree
• Graph

OOP Patterns

Languages

Machine Language
Assembly Language
C
C++
HTML
Java
Javascript
Jpython
Lex
Perl
PL/SQL	Oracles extensions to SQL to make it a general purpose programming language. Can create stored procedures inside the database
Python
SQL
XML
yacc/bison

Links

• Law of leaky Abstraction by Joel Sposky