Hadoop Concepts

Abstract

Object Oriented Programming, OOP, is the must influential paradigm of our time. This handout summarizes the most basic style, elements, and vocabulary of OOP that are common to all OOP languages. OOP languages can have weird features, but the basic ideas of OOP are pretty straightforward. Pre-OOP In a classical compiled language like Pascal or C, data-structures it is the programmer's duty to devise and enforce logical groupings of the data types and the functions that operate on them. The programmer can put related functions together in one file, but the grouping is just a convention and the compiler does not enforce it in a significant way. In C, a disciplined programmer can build code which is well structured. In OOP, much of the structure is a formal part of the language which makes it easier to get right. OOP In OOP, the logical arrangement of the code is changed. Instead of an informal arrangement of functions into different files, functionality is officially and tightly grouped with the type that it operates on. The OOP style groups all the operations together according to what they operate on — all the hash table operations are part of the hash table class, all the string operations are part of the string class. Put another way: if variables are nouns and functions are verbs, then OOP divides up everything around the nouns. Historically, there are many ways to structure code, but the OOP style has proved to work quite well. OOP yields a pretty obvious decomposition and facilitates code re-use. Most importantly, OOP keeps the role of client and implementor separate. Keeping the client side separate from the implementation side has many advantages. Don't just think "oh yeah, client separate from implementor — too basic for me." It's not sexy, but it's useful. OOP formalizes and enforces the separation. It no longer requires any special skill to keep them separate; the language makes it the most convenient way to proceed.