Problem F - Clean up that Web Page!

Many markup languages are in use for text formatting today, among them HTML and T_EX. A markup language provides a way to define formatted text in a readable ASCII file using tags, or codes, to denote formatting. In this problem, assume that our markup language, ACMML, uses tags in pairs, of the form <XXX+> and <XXX>, in that order, to surround text. When these tags surround text, they represent text formatted with whatever aftribute XXX indicates. For example, if BOL indicates bold and ITA indicates italic, the ACMML source text

This is how ACMML
encodes <BOL+>bol<ITA+>d<BOL>
 a<ITA>nd <ITA+>italic<ITA> text.

represents the formatted text

This is how ACMML
encodes bold
 and italic text.

The same formatted text can also be represented by this longer ACMML source text:

This is how ACM<BOL+><BOL>ML
encodes <BOL+>bol<ITA+>d
<ITA><ITA+><BOL> a<ITA>nd <ITA+><ITA+>ita<ITA>lic<ITA> text.

You will need to write a program to optimize ACMML source text.

INPUT: ACMML source text defining some formatted text.

OUPUT: ACMML source text defining the same formatted text, but containing the smallest possible number of pairs of tags.

• The codes (XXX above) within the tags are always three-letter sequences.
• The characters < and > appear in the input only as parts of tags. Those two characters cannot appear in formatted text.
• Format aftributes do modify white space, so that, for example, the source text
  <ITA+>i <ITA>A
  and
  <ITA+>i<ITA> A
  do not define the same formatted text.

Example:

Input File:

This is how ACM<BOL+><MXY+><BOL><MXY>ML
encodes <BOL+>bol<ITA+>d
<ITA><ITA+><BOL> a<ITA>nd <ITA+><ITA+>ita<ITA>lic<ITA> text.

Output File:

This is how ACMML
encodes <BOL+>bol<ITA+>d<BOL>
 a<ITA>nd <ITA+>italic<ITA> text.