Word count functions and main wrapper.

wordcounters.m

#import <Foundation/Foundation.h>

NSUInteger scannerWordCount(NSString* string) 
{
  NSScanner* scanner = [NSScanner scannerWithString:string];
  NSCharacterSet* ws = [NSCharacterSet whitespaceAndNewlineCharacterSet];
  NSUInteger words = 0;
  while ([scanner scanUpToCharactersFromSet:ws intoString:nil])
    ++words;
  return words;
}

NSUInteger regexWordCount(NSString* string)
{
  NSRegularExpression* regex = [NSRegularExpression regularExpressionWithPattern:@"\\w+" options:0 error:nil];
  return [regex numberOfMatchesInString:string options:0 range:NSMakeRange(0, [string length])];
}

NSUInteger componentsByStringWordCount(NSString* string)
{
  return [[string componentsSeparatedByString:@" "] count];
}

NSUInteger componentsByCharsWordCount(NSString* string)
{
  NSCharacterSet* ws = [NSCharacterSet whitespaceAndNewlineCharacterSet];
  return [[string componentsSeparatedByCharactersInSet:ws] count];
}

NSUInteger taggerWordCount(NSString* string)
{
  NSArray* schemes = [NSArray arrayWithObject:NSLinguisticTagSchemeTokenType];
  NSLinguisticTagger* tagger = [[NSLinguisticTagger alloc] initWithTagSchemes:schemes
                                                                      options:0];
  [tagger setString:string];
  __block NSUInteger words = 0;
  [tagger enumerateTagsInRange:NSMakeRange(0, [string length])
                        scheme:NSLinguisticTagSchemeTokenType
                       options:0
                    usingBlock:^(NSString* tag, NSRange token, NSRange sentence, BOOL *stop) {
    if ([tag isEqualTo: NSLinguisticTagWord]) ++words;
  }];
  return words;
}

int main(int argc, char *argv[]) {
  NSAutoreleasePool* pool = [[NSAutoreleasePool alloc] init];

  NSString* string = @"Peter  piper  picked  a  peck  of  pickled  pepper . No — really — he did!";

  printf("scanner (original)\t: %lu\n", scannerWordCount(string));
  printf("regular expression\t: %lu\n", regexWordCount(string));
  printf("components (string)\t: %lu\n", componentsByStringWordCount(string));
  printf("components (chars)\t: %lu\n", componentsByCharsWordCount(string));
  printf("linguistic tagger\t: %lu\n", taggerWordCount(string));

  [pool release];
}