🚀 go-pugleaf

scls_ss@uhura.cc.rochester.edu (Eric Scoles) writes:

>In <Jim_Johnson.0600@abcd.Houghton.MI.US> Jim_Johnson@abcd.Houghton.MI.US
>(Jim Johnson) writes: 
>>[...]
>>On a 286 machine or an 8086 with expanded memory, I think DRDOS 6.0 has
>>a bit more flexibility for memory useage than MSDOS 5.
>
>I'll just comment that my personal opinion is the opposite -- for most 286s,
>DRDOS is not as "good" -- you can only use UMBs if you have LEAP or NEAT
>chipsets, and you can't load that super disk cache high without EMS.

Since DRDOS supports the NEAT, LEAP, SCAT, and AMD chipsets in addition to 
large frame EMS for Upper Memory Block support and MSDOS will NOT support UMBs
on any 286 without the addition of a third-party memory manager, how can DRDOS
not be as good as MSDOS in that particular area (286 UMB support)?


>>I've heard that MSDOS 5.0 is a bit faster, but I haven't installed both
>>DOSes on the same machine running the same applications, so I really can't
>>say whether or not this is true.
>
>For some reason, the DRDOS update seems to boot a _lot_ faster than the
>[files] it replaces.

One interesting test is to use Manifest's timer interupt latency test to
compare QEMM, MSDOS EMM386.EXE, and DRDOS EMM386.SYS on the same 386 system.
The relative performance of MSDOS was some 20 times slower than QEMM. (DRDOS
was 0.25 times slower.)  This is why QEMM is still a viable product.

The best thing about having two competing products is that the customer comes
into the equation now.  Even if someone were to dispute that DRDOS is an 
emormously better value, they would have to admit that MSDOS is better now
BECAUSE of DRDOS.  Everyone remembers how Gordon Letwin of Microsoft came
before the net two years ago to ask for input on desired features for a 
proposed new version of MSDOS only after DRDOS 5.0 was first publicly shown. 
 

(Note: I work for the company that makes DRDOS.)

--
Jimmy Liberato   liberato@dri.com
                 ...uunet!drivax!liberato