Mail Archives: opendos/2000/01/25/21:47:16
On Tue, 25 Jan 2000 07:50:18 +0100 (MET) Bernie
<bernie AT mbox302 DOT swipnet DOT se> writes:
> Bruce Morgen wrote:
> (Please note that I left out most of the parts where I could say
> something
> like "I agree.")
>
> >That API exists only after
> >the GUI is loaded and running.
>
> Yes, that's correct.
>
> >> IMHO those two are all that is needed for DOS (+ command.com or
> >> something like it of course).
> >
> >Actually, only IO.SYS and a
> >shell like COMMAND.COM are
> >needed -- from DOS 7 on,
> >MSDOS.SYS is a text file
> >used by the GUI loader
> >started by WIN.COM.
>
> You are very correct, sorry my misstake.
No problem, we live to
serve. :-)
>
> >> In Win9x the computer boots DOS
> >
> >Yes!
> >
> >> (with 32bit API) and
> >
> >No -- this API doesn't exist
> >until the GUI is running!
>
> Hmm... Sorry but no. If that was the case then I couldn't be able to
> start
> win.com at all on my computer since the 32bit APIs have a very nasty
> bug in
> them that makes my AMD K6-2 400 stop ("IOS error" or somthing like
> that). I have now installed the bugfix.
Sorry, you're logic here
eludes me. Are you implying
that the "bugfix" is to a
DOS component and not some
part of the GUI?
>
> >One can confirm the absence
> >of the 32-bit API by the
> >fact that no long filenames
> >are visible unless the GUI
> >is running!
>
> Doesn't the 32bit API provide anything else but LFN and FAT32
> support?
Yes, I'm just using those as
indicators of the API's
presence, not the (vast, I'm
sure) extent of its services.
>
> >Wrong. These are DOS 7 files
> >and can't run without a
> >running GUI if you can do
> >without long filename support.
>
> I must say I have never tried... I'll do a test later today if I
> remember.
> (setver.exe might be needed)
Nope. XCOPY, FC, FORMAT, etc.
in the WINDOWS\COMMAND
directory will run fine in
"Safe Mode Command Prompt
Only," which is as raw a DOS 7
as you can get!
>
> >> * (more or less) impossible to install hardware into it
> >
> >Sorry, here we disagree. If
> >you do things in the correct
> >order, Win95's hardware
> >detection is really quite
> >reliable given the correct
> >driver(s).
>
> Really? I now have my NIC, SB16 and TV-Card on the same IRQ (5)
> because
> that's the only way I've found (so far) that Windows accept them
> (impossible to find the NIC otherwise - but I knew it was working).
> And
> neither is now recognized as the things they are so the drivers
> don't work (luckily things seems to work anyway).
Your hardware doesn't seem
to support enough IRQ
alternatives, can't blame
that on Windoze. It have a
similarly crowded card
collection -- NIC, modem,
two active serial ports,
IRQ/DMA-hungry sound card,
SCSI controller, Sony
proprietary CD-ROM interface
-- with no conflicts at all.
>
> >> * slows down your computer so much
> >
> >Again we disagree, on my
> >pre-Pentium hardware Win95 is
> >no slower overall than Win3.11.
>
> Hmm.. Well it's slower on all my computers (486, Pentium, Pentium
> MMX and
> AMD K6) then DOS.
Sure, but DOS isn't really
an operating system, it's a
sort of extended monitor
with disk access -- warm-
over CP/M with a few Unix-
like enhancements. It has
virtually no overhead
because (outside of things
like device drivers and
memory managers) it's
pretty much idle until its
called on to do something.
Multitaskers like Windoze
and UNIX are active all the
time and are much more CPU-
intensive -- of course they
actually retain control of
the hardware, whereas
single-task, non-reentrant
stuff like DOS pretty much
steps aside and lets the
app of the moment take
control.
> Although Win 3.x multitasking has always worked better
> then Win9x's (I know I'm relativly alone on that)
Yes, you are. Win95 has
been both (a little) less
prone to crash and (a lot)
better at crash recovery
than Win3.11 here.
> I have no idea if Win 3.x
> or Win9x is fastest.
It's Win95 by a nose here,
especially when comparing
32-bit vs. 16-bit versions
of the same app.
>
> >It does take up more RAM, but I
> >have a fairly fast SCSI disk
> >subsystem and disk swapping is
> >quite fast -- and much less
> >crash-prone that is is under
> >Win3.11.
>
> It's not often I get crashes in 3.x - 9x (or NT for that matter)
> crashed on
> a daily basis for me (even if I only used them for a few minutes).
I found Win3.11 crashed at
least once a day, usually
more. Win95 rarely crashes
more than once a day and
generally doesn't require a
reset button cold boot like
Win3.11 often did. Faint
praise for M$, I know. :-)
> Tip:
> Keep clear of the right mouse button in 9x and NT.
Thanks, I've never noticed
that one, now I'll be alert
for it!
>
> >> * way to big (or "demanding on the hardware")
> >
> >Win95 works OK in 16, 24, or
> >32 MB. I can't say the same
> >for Win98 or NT.
>
> But I'm not comparing with those OS but with DOS.
That's an apples vs, oranges
comparison, but I take your
point.
>
> Anyway, it's important (IMHO) that we point out the things that are
> incorrect in Windows and not just poor out random errors that are
> results of incorrect installations etc.
As far as I know, my Windoze
install is clean and conflict-
free. It still crashes, but
perhaps there's an intermittent
SIMM or cache chip -- it's hard
to know when you run a
homebrew system built largely
from old/spare parts! :-)
> I think that if one hears one thing in an arguement that is easily
> proved
> wrong then the entire discussion is believed to be wrong by many
> people.
You certainly could be right
there, Bernie. Thanks for
writing -- and apologies to all
for wandering so far off-topic!
________________________________________________________________
YOU'RE PAYING TOO MUCH FOR THE INTERNET!
Juno now offers FREE Internet Access!
Try it today - there's no risk! For your FREE software, visit:
http://dl.www.juno.com/get/tagj.
- Raw text -