Here's something that sort of does what you want:

#!/usr/bin/perl -w
use strict;
use Text::Balanced ('extract_tagged');

my $text = "this is a test <b>for</b> tags! \n";

my ($leading,$extracted, $remainder);
$leading = substr $text, 0, index($text,'<'), '';
($extracted, $remainder) = extract_tagged($text);

printf "leading: %s$/", $leading;
printf "extracted: %s$/", $extracted;
printf "remainder: %s$/", $remainder;

exit 0;
[download]

That can be iterated over both $extracted and $remainder to parse html. HTML::Parser is probably a better bet, particularly (given your reply to jcwren) with its friend HTML::LinkExtor.

Update: Here's code that does what you were trying more exactly. Text::Balanced honors pos:

#!/usr/bin/perl -w
use strict;
use Text::Balanced ('extract_tagged');

my $text = "this is a test <B>for</B> tags! \n";
pos($text) = index $text, '<';

my $extracted = extract_tagged($text);

printf "extracted: %s$/", $extracted;
printf "remainder: %s$/", $text;
exit 0;
[download]

Note scalar context for &extract_tagged.

After Compline,
Zaxo