🚀 go-pugleaf

I've noticed a few (very few) differences between MS-DOS and 4DOS/DRDOS command
interpreters too.  One affects the way you can check for a directory existing
by testing if the file dirname\.. exists, but I don't think the difference Timo
Salmi mentioned comes from the different COMMAND.COM - rather what the
operating system considers to the the reason for error; IORESULT from a RESET
in Turbo Pascal shouldn't have anything to do with what command interpreter is 
running... at least as far as I can see from /pc/pd2/post10.zip.

But the differences between them isn't great (not when you look at differences
from one version of MSDOS to the next), and (as somebody said) if it breaks a
program then the program was probably going to break anyway some day.

That sounds a bit harsh towards the programs, and biased in favour of
replacements for command.com.  Well, I expect MSDOS will change a lot in the
future, as well as a lot of new look-alikes, which is a good thing. I'd like to
see a list of the "undocumented" tricks that people reasonably expect of DOS
and command interpreters, which programmers stick to AND writers of new DOSes
and command shells take notice of. So I wouldn't like DOS to be limited by lots
of old programs expecting unreasonable compatibility, nor would I like to see
programs have to restrict themselves to just the documented DOS calls.

By the way, In article <1992Aug22.181032.11678@ctr.columbia.edu>, kibirev@csa.bu.edu (oleg kibirev) 
wrote:
> ... Hmmm, on my computer it took about the same time for DRDOS to display
> A:\> prompt. Or to traverse the search path. At least without SUPERPCK
> DRDOS 6.0 disk performance is _horrible_ compared to MSDOS 5.0.

This a well-known bug/limitation/whatever that has been discussed before. Some
DR bloke, I think, said something about DRDOS 6 (not 5?) taking care not to
directly transfer multiple disk sectors into high or upper memory.  I can't
remember the details, but the result is that one sector is read per revolution
of the disk! The answer is to use SuperPCK with the /T option, and I think that
it is only a problem with some configurations, and that using DRIVER=... also
fixes it.


Mark Aitchison, University of Canterbury, New Zealand.