This is NOT the UCSD p-System filer, it's one belonging to a derived version of it that Apple briefly used on the Apple II. To give you a little context, this was version II of the p-System (the Roman numerals denoted architecture versions). The one released by SofTech Microsystems, on the IBM and the Sage and many others, was version IV, and it was a major overhaul. (Status of segments regularised, almost no limit on the number of segments in a system, all units swappable, dynamic linking, instruction set even more compact than before, procedural parameters, conformant arrays, semaphores, processes, etc. etc.) Including using 80 columns on the screens! (Actually you could change that in SYSTEM.CONFIG if you needed to, as seems clearly to have been done with this example). The filer used full names in its menus when it had the screen width to do so. Which was most or all of the time after those very early screens, unless, as I say, you did something to SYSTEM.CONFIG . As for making DOS look good, it made it look dreadful. I still remember the pain of migrating to Turbo Pascal, which was the BEST that DOS had to offer. Tiny file names, NO file typing (file extensions are NOT file types), terribly slow floppies, file volumes WITHOUT NAMES (which continues to this day), every program needing all its code linked into it statically, hence several copies of the same code on the same floppy -- etc etc. (DOS's only advantages: more flexible number of entries in a directory, and borrowing a poor man's version of directory trees from UNIX. The p-System only allowed 77 entries in a directory -- that was from a time when the DEC PDP-11 was popular -- and had only 1 level of "subdirectory", which was an actual subvolume file, of fixed size. And DOS was to acquire networking, whereas work on the p-System seems to have suspended before that could happen.) Oh, and the p-System had powerful wildcards, including use for mass name changes. DOS's wildcards were always weaker, and buggy. And I really did like being able to compile on a 68000 machine and run the compiled code on an 8086. Much faster turnaround time. (I once did a project where we were supposed to use the then-Microsoft Pascal on XT's for compiling -- all in native code of course. We stayed in UCSD Pascal and compiled within the p-System because it was FASTER!) I trust this misimpression is now corrected. A. Milne