On 23 Aug 2012 at 12:21, Allison wrote:
It's biggest limitation for me was the flat
filesystem. It made
large disks awkward (aside from the 8mb limit CP/M2) for lots of
files. That and extended memory support in later versions was
not transparent to the aps.
Going to a different file system was about the smartest thing that
(SCP/Microsoft) ever did. The rest of the MS-DOS and CP/M API is
pretty weak, but that filesystem made a huge difference. People may
gripe about Microsoft "stealing" from DRI, but if I were in the
driver's seat at IBM, there would be no questions as to which system
was superior. And what is CP/M or MS-DOS but a file management
facility?
There was TurboDOS, but it wasn't even close to being an open system.
We did our own OS, complete with ISAM support and loadable device
drivers. Again, the CP/M internal interface didn't envision that
people would want to support other devices than a simple console,
printer, reader and punch. As such stuff was available on mainframe
OS, I can't say that DRI was forward-looking.
ZCPR was a big CLI improvment, but DRI evidently didn't think it was
worth emulating.
--Chuck