Christian Neukirchen wrote:
> "William James" <w_a_x_man / yahoo.com> writes:
> 
>> class String
>>   def xtag(s)
>>     scan( %r! ( < #{s} [^>]* > ) ( .*? )  </ #{s} > !mx )
>>   end
>> end
>>
>> gets(nil).xtag("biblioentry").each { |tag,data|
>>   if data.xtag("pubdate")[0][1] > "1984"
>>     print tag, data, "\n"
>>   end
>> }
> 
> I hope you are joking...
> 

Actually, in real-world usage, Mark Pilgrim's Python Feed Parser[0] 
falls back to regular expressions to get the data required if the XML is 
not well-formed.

Admittedly this is a real problem for RSS hackers, less so with other 
XML messages, but the approach does have merit if (a) you can't 
guarantee well-formedness and (b) you absolutely have to have the data.

-dave

[0] http://feedparser.org/