OE-3.4 Collections Framework & I/O

Work through these

• Overview of the collection framework

  The library's answer to storing groups of things, taken as a map before the parts. Knowing the shape of the framework is what lets you choose rather than reach for the same class every time.

• ArrayList, LinkedList, HashSet, LinkedHashSet, TreeSet

  The concrete containers, and the reason there are several: they trade insertion cost, lookup cost and ordering differently. Choosing one is a performance decision, not a style one.

• Collection, List, Set and SortedSet interfaces

  The contracts those classes implement, which is the level you should usually write against. Programming to the interface is the habit this item exists to build.

• Accessing a collection via an iterator; storing user-defined classes

  Walking a collection, and what your own class must provide before it can live in one. Equality and hashing are the requirements people forget, with results that look like the collection is broken.

• Map interfaces and classes; using a comparator

  Key-value storage, and how to impose an order the type does not have on its own. Comparators are the piece that makes sorting arbitrary objects possible.

• Legacy classes: Vector, Hashtable, the Enumeration interface

  The older classes that predate the framework, which you will meet in existing code. Know them well enough to read, and prefer the modern equivalents when writing.

• Reading user input with the Scanner class

  Getting input from a person, which is the first thing most programs need. Small, and it is where the exception handling from the previous topic first earns its keep.

• Stream classes: InputStream, OutputStream, FileInputStream, FileOutputStream

  Reading and writing bytes, and the layered design the library uses. The layering looks elaborate until you need to add compression or buffering, at which point it pays.

• Reader and Writer, FileReader, FileWriter and the File class

  The character-oriented half, for text rather than bytes. Using the byte classes on text is a real bug that produces mangled output rather than an error, so the distinction matters.