Concept for a PHP String class

class-concept.php

<?php
//============================================================
// CONSTANTS

// direction (for strpos etc)
String::FORWARD;
String::REVERSE;

// case
String::CASE_LOWER;
String::CASE_UPPER;
String::CASE_TITLE;
String::CASE_SENTENCE;

// position (for trim, pad etc)
String::POSITION_RIGHT;
String::POSITION_LEFT;
String::POSITION_BOTH;

//============================================================
// METHODS

// define
$str = new String('Example');
// or
$str = string('Example');

// case changes
$str->case(String::CASE_LOWER); // String('example')

// get a substring
$str->part(2, 3); // String('amp')

// get position of a substring
$str->position_of('a', String::FORWARD); // Integer(2)

// split a string
$str->split('m'); // Array('Exa', 'ple') (also second argument is max number of chunks to return)

// return the character at a position
$str->character(0); // String('E')

// get the string length
$str->length(); // Integer(7)

// trim characters from a string
$str->trim('e', String::POSITION_BOTH); // String('xampl')

// pad a string
$str->pad(10, '-', String::POSITION_RIGHT); // String('Example---')

// perform a replacement
$str->replace('Ex', 'S'); // String('Sample')

// ..etc

//============================================================
// CHAINING

// 'Example' to 'examples'
$str = string('Example')
	->case(String::CASE_LOWER)
	->pad(8, 's')
	;

// if Integer/Array classes etc. were created...
// 'Hello World!' to 'World, Hello'.
$str = string('Hello World!')
	->split(' ') // now an array
	->reverse()
	->join(' ') // now a string again
	->replace('!', ',')
	;

//============================================================
?>

Is this ridiculous? Am I over-thinking? Would this cause your scripts to grind to an appallingly slow speed? I'm genuinely interested – I think it'd speed up the way I work and make my code more readable... What are your thoughts?

I can see it being useful to have up your sleeve - esp when writing little apps that chug away at large amounts of data (e.g. parsing/cleansing CSV files) – have you done any comparisons against native string manipulation yet to see how it fares? If the cost/benefit (speed of execution vs speed of coding) balances, could be onto a good un!