[Koha] Unimarc - usage of UDC in build_browser_and_cloud.pl

[julionogueira at sapo.pt]

Hello everyone,
The hierarchical search is, in my opinion, an extremely useful feature.
I activated it by following the steps which José Santos indicates on  
http://koha.pt/forum/index.php?topic=5.0.
The searches return good results when systematic notations contain up  
to three numbers because, as we know, build_browser_and_cloud.pl  
script defaults to three numbers.
But if we take a look at the Universal Decimal Classification main  
table (Portuguese abridged edition), we see different types of  
punctuation between the second and third numbers, or even between the  
first and second numbers.
Additionally, there are many notations that can have a dozen or more  
characters consisting of numbers and punctuation marks.
Some typical examples:
"2-1” => “Theory and philosophy of religion. Nature of religion.  
Phenomenon of religion”,
"34.01” => “Nature of law. Objectives of law",
"616-073.65/.66” => “Thermometry. Temperature measurement. Calorimetry”,
"81'373” => “Lexicology. Onomastics. Toponymy”,
"821.133.1(569.3)” => “Lebanese literature in French”,
"94(100)\"1939/1945\”" => “World War II”, # or 94(100)“1939/1945" with  
a pair of Perl backslash escape character before the double quotes to  
differentiate script and text.
This obviously means we cannot get these notations in search results  
and these constraints call into question the integrity of the  
hierarchical search.
What could be changed in the script to take into account punctuation  
marks and larger notations?
Many thanks in advance and regards from
Júlio
(librarian)