(update: the later example, Re: module for HTML XML conversion, is much better than this one.)

This is pretty specialized and inlines more than it should. You have to write a few lines of code for every part of your spec. So it is not the best if you have a need for generalized transforms but if you have an exact transform to do that won't change much, this can be great. XML::LibXML.

use warnings;
use strict;
use XML::LibXML;

my $parser = XML::LibXML->new();
$parser->keep_blanks(0);

# $parser->recover_silently(1); # <-- for bad HTML

my $doc = $parser->parse_fh(\*DATA); # If you have well balanced snipp
+et like below

# my $doc = $parser->parse_html_fh(\*DATA); # If you have HTML.

my $xml = XML::LibXML::Document->new();
$xml->setDocumentElement( $xml->createElement("engines") );

for my $spec_row ( $doc->findnodes('//div[@class="spec_row"]') )
{
    my $engine = $xml->createElement("engine");
    $xml->getDocumentElement->appendChild($engine);

    my $type = $xml->createElement("type");
    ( my $text = $spec_row->firstChild->textContent ) =~ s/\A\s+|\s+\z
+//g;
    $type->appendChild( $xml->createTextNode($text) );
    $engine->appendChild( $type );

    my $builder = $xml->createElement("builder");
    ( $text = $spec_row->lastChild->textContent ) =~ s/\A\s+|\s+\z//g;
    $builder->appendChild( $xml->createTextNode($text) );
    $engine->appendChild( $builder );
}

print $xml->serialize(1);

__DATA__
<div class="spec_section">
              <div class="spec_row">
                <div class="spec_row_header orange_text_1">
                  Engine                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Type                </div>
                <div class="spec_row_right">
                  Diesel                 </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Builder                </div>
                <div class="spec_row_right">
                  MTU                 </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Model                </div>
                <div class="spec_row_right">
                  2x 16V396TB94                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Power                </div>
                <div class="spec_row_right">
                  2561kw / 3480hp                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Total Power                </div>
                <div class="spec_row_right">
                  5121kw / 6960hp                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Engine Propulsion                </div>
                <div class="spec_row_right">
                  Twin Screws                 </div>
              </div>
</div>
[download]

my $doc = $parser->parse_fh(\*DATA); # If you have well balanced snippet like below

Is there any reason to not use parse_html_fh to parse HTML? All of the following are well balanced (or impossible to balance) but invalid XML:

<option selected ...>...</option>

<font color=red>...</font>

<br>
<img ...>
<meta ...>

&nbsp;    (without XML bits to define it)
[download]

Encoding is also handled differently.

No, I think it's a good idea. But the sample input was well balanced so I used it and included the other in (as you point out, the likely) case it wouldn't fly.

The sample also fails to specially account for the first <div/> which contains the "column" names. We could probably do something automatic with that instead of hard coding the column parsing. The OP might have thought of that already. I didn't until looking it over again now. If I have a few minutes I'll revisit it with that approach.

That's very kind of you to do that - I can definitely work on that approach. Much appreciated...

From what I've seen, XML::Twig is great at transforming XML. If you could transform your HTML to XHTML somehow, then you could use XML::Twig to transform the XHTML into the XML subformat you desire.

HTML::TreeBuilder + ->as_XML

I think my first example was not getting very close to what you actually want. Try this instead. It's also a bit weird, tricky, and not robust but the output seems right this time and it's semi-automatic for finding key, value pairs.

use warnings;
use strict;
use XML::LibXML;

sub snipName  { s/\s+//g and $_ = lcfirst($_) for @_ };
sub snipValue { s/\A\s+|\s+\z//g for @_ };

my $parser = XML::LibXML->new();
$parser->keep_blanks(0);
# $parser->recover_silently(1); # Might need, might not.
my $doc = $parser->parse_html_fh(\*DATA);

my $xml = XML::LibXML::Document->new();

my $root_name = [ $doc->findnodes('//div[@class="spec_section"]//div[c
+ontains(@class,"spec_row_header")]') ]->[0]->textContent;
snipName($root_name);
$xml->setDocumentElement( $xml->createElement($root_name) );

for my $col ( $doc->findnodes('//div[@class="spec_row_left"]') )
{
    snipName( my $name = $col->textContent );
    my $tag = $xml->createElement($name);
    my $value = [ $col->parentNode->findnodes('div[@class="spec_row_ri
+ght"]') ]->[0]->textContent;
    snipValue( $value );
    $tag->appendChild( $xml->createTextNode( $value ) );
    $xml->getDocumentElement->appendChild($tag);
}

print $xml->serialize(1);

__DATA__
<div class="spec_section">
              <div class="spec_row">
                <div class="spec_row_header orange_text_1">
                  Engine                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Type                </div>
                <div class="spec_row_right">
                  Diesel                 </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Builder                </div>
                <div class="spec_row_right">
                  MTU                 </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Model                </div>
                <div class="spec_row_right">
                  2x 16V396TB94                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Power                </div>
                <div class="spec_row_right">
                  2561kw / 3480hp                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Total Power                </div>
                <div class="spec_row_right">
                  5121kw / 6960hp                </div>
              </div>
              <div class="spec_row">
                <div class="spec_row_left">
                  Engine Propulsion                </div>
                <div class="spec_row_right">
                  Twin Screws                 </div>
              </div>
</div>
[download]

Which yields-

<?xml version="1.0"?>
<engine>
  <type>Diesel</type>
  <builder>MTU</builder>
  <model>2x 16V396TB94</model>
  <power>2561kw / 3480hp</power>
  <totalPower>5121kw / 6960hp</totalPower>
  <enginePropulsion>Twin Screws</enginePropulsion>
</engine>
[download]

FWIW I am doing this with HTML Tidy ... via the shell http://www.ibm.com/developerworks/library/x-tiptidy.html

Javascript DOM