in reply to Data mining

Looks good to me. (don't forget the closing sites tag.)

A small consideration, however, is this: 

<site>
   <name>Ralinga</name>
   <url>http://www.ralinga.lt/</url>
   ...
</site>

[download]
versus this: 
<site name="Ralinga" url="http://www.ralinga.lt/">
   ...
</site>

[download]
It really is only a matter of personal choice, though. I find it very helpful to shove my XML through XML::Simple (with and without forcearray on) and then shove the resulting data structure through Data::Dumper to see the differences.

UPDATE:
After some reflection, i am curious why you didn't break this line up: 

<extract type="info">TD class=kain width=500</extract>

# which would yield a data structure similar to
'extract' => [
   {
      'content' => 'TD class=kain width=500',
      'type' => 'info'
   }
]

[download]
Wrapping the content in a simple tag would do the job of breaking up the atoms for you: 
<extract type="info"><TD class="kain" width="500"/></extract>

# yields something like:
'extract' => [
   {
      'TD' => {
         'class' => 'kain',
         'width' => '500',
      },
      'type' => 'info'
   }
]

[download]
Let the XML parser do the work. ;)

jeffa

L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)

In Section Seekers of Perl Wisdom