As I understand things, the ROM BIOS on a PC reads the first track of
the disk to load the boot strap loader, or something to that effect.
Could an Intel box running Win32 realistically NOT use IO.SYS and
MSDOS.SYS, AND still allow one to 'boot the previous DOS version'? In
other words, could Microsoft have NOT adapted these files to start
Windows, without losing anything? I read and enjoyed the first several
chapters of Schulman's Undocumented Windows 95 (and recommend it to
those of us who enjoy understanding those inner workings; if you loved
The Design and Evolution of  C++...), and it seems to me that this is at
the heart of the contention, closely followed by the existence of
several 32 bit goodies in WFWs that deserved more than a .01 change to
the version number (I smell marketing on this decision). At this point,
it becomes 'a matter of interpretation', although I expect compelling
arguments could be made on either side for what is and is not DOS, 100%,
99 and 44/100ths %, or otherwise. Schulman does ask the question if one
would have wanted MS to jettison the tried and true SYS files in favor
of something else. Personally, I am more comfortable with their sticking
with the known quantity.

Now, the JVM, there's a thing on a thing for you! :-)

Opinions are mine, copyrights are Microsoft's and Ivory's.
>