HTML markup language consists of text interspersed with "tags" which provide information about how the text is to be displayed. Each tag is encapsulated between `<' and `>'. For instance, `<p>' indicates the start of a new paragraph and `</p>' indicates the end of a paragraph. <ol> starts a numbered list, <li> starts an item of the list, and </ol> indicates end of the list.

When editing HTML source text containing all these tags, it is very helpful if the source is indented in a systematic way, very similar to the indentation we use with C++ programs. The rules of the indentation we want are very simple. Divide the text into segments called "tokens". A token is either an html tag or a portion of text containing no tags and starting from the end of a previous token to the end of a line or to a tag.

• Call a section between a start tag and matching end tag a "block". Indent the block two spaces more than the start tag and end tag themselves.
• There may be several start tags with the same end tag. Indent them equally. See example below.
• After a start tag and until the matching end tag, indent everything more than the start tag is indented.
• Indent a start tag just as much as previous text, but more than the previous start tag. (or, if the previous start tag is the same as this one, use the same indentation).
• Indent a text line more than the previous start tag, all text lines with in the same tagged region being indented the same.
• Indent an end tag (with the backslash, such as `') as much as the corresponding previous start tag. If there is no corresponding start tag, revert to zero indentation.
• As an operational definition of what all this means, your indentation pattern on any input text should be the same as that produced by the program `mystery' in the project directory.

<ol> <li> first item <li> second item </ol>

<ol>
  <li> 
    first item 
  <li> 
    second item
</ol>

1. first item
2. second item