I think you guys are going down two different paths, at least for the moment. Which is not to say Jonathan couldn't change directions. (I'm not making a judgement her about the direction anyone *should* go, I'm just trying to shed some light ... :-) 1. Jonathan is trying to merge Bert's Winfract with the latest DOS sources using the same compiler we use for the medium model DOS fractint. This approach is theoretically possible because Winfract is a Windows 3.1 program which uses the same medium memory model as the DOS fractint. 2. Paul is using a newer compiler and will therefore have to eliminate all vestiges of the medium memory model (mostly the assembler code.) Of course Paul has been down this path before with his own program which is derived from Fractint. The almost certain reason Jonathan is getting protection faults is that fractint is way too big to fit into the medium memory model without overlays. The current DOS version works with the medium model only because we gave a whole lot of thought and shoe-horned it in. Bert had to do something similar with Winfract, but it was easier then because Fractint was smaller. But when Jonathan merged the current sources, Bert's overlay scheme probably broke. As Rich pointed out, if Paul or anyone wants to make as literal a port of Fractint as possible to Windows with a newer compiler, the Xfractint version is a better starting point because it has already got the flat memory model - but there's a lot of work stripping out the Xwindows GUI and replacing it with a Windows API. Neverless I was excited to see Jonathan attempting to merge Winfract with the current DOS version. If you look up Fractint in wikipedia, it is cited as one of the oldest open source programs still maintained (thanks to Jonathan). The other two mentioned are Emacs and NetHack. Tim