I'm working on a small project to parse SAP's evil idoc files, and convert them into UPS's equally evil XML file format for their World Ship product.
It's not a big project and it's ideally suited to Perl. I was amazed by how little there is out there for working with idoc/edi and Perl. It seems some else did the same searching that I did and also drew a blank, and asked a question on Perlmonks a few hours after I had a good look there.
I started hacking my code for work, the framework is dead easy, even if I don't yet have a parsing engine for the edi/idoc file format. I realised that I was reusing simple logging code from another similar project. Rather than cut'n'paste it, I decided to modularise is, and Log::Trivial went up to CPAN on Friday. Rather than do the work at work for work, I did it at home in my own time so that I can release it on CPAN. I'll use it and extend it at work, but putting it on CPAN in my name first will avoid any problems with work not understanding the GPL and open source.
There is has been no problem in using GPLed code at work, and even contributing back. Managers can see the advantage of having code on CPAN tweaked in away we want rather than having to maintain our own fork. However when a colleage tried to releace something new done specifically for work at work, he met sufficient resitance to not release the code.
