On Thursday, Jul 24, 2003, at 12:54 Canada/Eastern, TheMacintoshLady wrote: > It would have been nice if there could have been just one location like > there used to be and one method of installing. Definitely not user > friendly! Not at all. In fact, it would have been very bad indeed. A corrupt font could have crashed the entire system (as it used to happen in 9). My wife's fonts could not have been kept separate from my own fonts -- doesn't any modern marriage have enough stresses already? If an application had required a special font (e.g., for terminal emulation) of no earthly use to anyone else, it could not have been kept separate from the other fonts. An office manager would have had to manage each Mac on her network individually to make sure they all had a consistent complement of fonts. A casual user could have wiped out fonts used by the system to draw the menus and dialogue boxes, with dire consequences. There are really two ways you can go about it. You can think of X as a souped-up version of 9, and try to change your ways as little as possible, and run into all sorts of pickles, all the while complaining about this which was poorly thought-out and that which is not user-friendly. Or you can think of X as a completely new operating system, with some similarities to OS 9 -- and some similarities to Windows, for that matter -- and learn from scratch to do things the OS X way. The OS X font locations scheme makes little sense for a antiquated single-user OS, as was the 'Classic' Mac, but it makes perfect sense for a networked multi-user environment. The fly in the ointment is the Classic fonts folder, but that can't be helped -- it's required for legacy applications. However, Apple was amiss in not providing the user with a method of controlling fonts. The scheme is perfectly logical, but the user should have had the possibility of seeing at a glance which font was where and whether he could disable it -- somewhat in the fashion of Stephane Sudre's Diablotin. By all accounts, Panther's Font Book goes some way towards providing this, but it should have been implemented at least in v10.1, not in v10.3. f