Odds & Ends: Radeon PCI card and VGA connector; NAV 7; Final Cut Pro 2 real-time editing and video cards; Virtual PC and i

Radeon PCI card VGA connector issue Robert Hancock writes: "The VGA connector in the ATI Radeon PCI card is too close (by about 4 mm) to one end of the back plate so that when you try to fit the card in a tower-type beige Mac like the 8500, the edge of the Mac opening strikes the VGA connector mounting nuts preventing the card seating properly and also preventing connection of the VGA plug from the monitor. The DVI connector is not a problem - it's in the center of the back plate. The simple answer is to take a very sharp modeling knife and carefully cut back the edge of the mac opening by about 5 mm. There's a metal RF shield behind the plastic but once the plastic is cut back, the shield does not cause any problems. You could also try removing the offending VGA mounting nut from the Radeon card but I still couldn't get the monitor connector to make a good connection."

NAV 7 and scanning mounted disk images Frans van der Geest claims that Norton AntiVirus 7.x does not scan for viruses in mounted disk image files, even though Norton AntiVirus 6.x did so. We have not confirmed this.

Real-time editing with Final Cut Pro 2: Beyond Matrox Regarding our previous comment that real-time editing in Final Cut Pro 2 requires a Matrox RTMac board, Mark Boszko replies: "The real-time editing in FCP2 also works with the Pinnacle TARGA CinéWave card, providing real-time effects and editing with uncompressed video. The real-time function should also work with the Digital Voodoo D1 Desktop, also uncompressed, and the Aurora Igniter card (high quality M-JPEG, or uncompressed with option), though both companies still have yet to return comment, pending the actual release date of FCP2 and QT5."

Virtual PC versions 3 and 4 do invalid IEEE mathematics Bill Gletsos writes: "On a real Pentium PC running Windows 95 when a floating point number is divided by zero an exception is raised. Also if you do the SQRT of a negative number an exception is raised. The application program can then trap the exception. However in Virtual PC the division by zero returns INF as the result and no exception is generated. For the SQRT of a negative number -NAN is returned and again no exception is returned. Naturally this misbehavior causes the application programs to malfunction. It would appear that this is due to the fact that on the PowerPC the results INF and -NAN are the results returned by IEEE math but they are not the results returned by real Pentium PC's. Connectix should force the corresponding PC exception to occur when one of the results INF , -INF, NAN, or -NAN is returned by the PowerPC IEEE maths unit. This problem was discovered whilst testing a Delphi 5 Application under Virtual PC."