To answer your question: the original data isn't unicode because it isn't unicode. It just isn't. This database and application I am working on is 7 years old, and decisions were made back then for whatever reasone they were made. I want to work in the legacy encoding for the same reason anyone wants legacy support: to support legacy code and databases. It seems odd that to use XML in Perl that I cannot use that with a legacy character set. XML supports the legacy character set. So why doesn't XML::Simple support legacy encodings that XML itself supports? Honestly, I don't get it. It's either some engineers on a high horse saying all apps should be re-written for unicode for the common good to keep the software gods happy. Or there is a real practical reason. But XML:Simple is the only piece of software in my stack that is making it difficult to work with my legacy code and legacy database and legacy encodings. Perl works just fine with legacy encodings. XML works just fine. MySQL works just fine. HTML works just fine. Only XML::Simple doesn't. I know the benefit of unicode when combining data source docs in different encodings. That is great for those applications that need that. But my legacy application doesnt' need to do this. It just needs CP1252 and that's all it needs. So it is the case that the XML::Simple authors demand I rewrite all my legacy code and database to support unicode in order to use XML::Simple? Is that the only option? I would really rather not change all that legacy code and data.