I am working now on a project which implies lots and lots of xsl transformations. We have about 100-150 xslt files, and one of them are growing in length each day. In the last monthes I start wondering what will happen when the current developers will go away and we will have new colegues. How are they supposed to understand some XSLT of about 2000-3000 lines?
I start doing some XSLT refactoring and the feeling about this issue changed a little bit, but still i am asking myself if somebody discovered a more generic maintenability techinic.