In the near future, presumably, I'll have to read and process complex xlsx documents containing a lots of formulas. Jexcelapi can't even open those kinds documents, so I turned to Apache POI . Unfortunately, POI also have problems with some formulas becase support of this field is not fully laboured. [ Update 2010.03.10: After some experimentations I found that Apache POI can read and process new Excel ( xlsx ) documents including calculating very complex formulas: Basically, cell.getStringCellValue() gives the evaluated outcome. This way, I didn't have to hack POI. That's great!] I choosed to start two methods: Knowing exact static structure of the Excel template document, I do problematic calculations in the reader program itself, eliminating POI's formula support. I study Apache POI's internal structure and try to make up the missing functionality. Unfortunately, POI isn't amendable in a non-invasive way with registering components at certain extensi