Hi Christian,

You wrote to Martin Foster on 29.10.03 about "new beta Version
OpenXP (3.8.13)":

CB>>> Do I get it right that you're testing the DOS executable on
CB>>> Windows? Is it just for testing?

MF>> Yes and yes.

CB> OK.

Just to try and explain the "method in the madness" of Martin's  
and my testing approach:

a.  Many/most(?) of the "newer generation" of PC users, tend to  
be Windows users of some sort.  To those folks a DOS program  
means using a Windows DOS box.  So we really must check that a  
DOS executable will work in that environment.

b.  Then there are the "pure" DOS users, a shrinking group as  
you'll appreciate.  We need to know which versions of DOS the  
executable will work with, or display problems.  In general those  
"pure" DOS users will be using MS-DOS - probably v6.22, PC-DOS -  
probably v6 or v7, DR-DOS - mainly v7.03, or perhaps something  
like FreeDOS.   If an executable works OK under MS-DOS v6.22, the  
chances are that it will be OK under PC-DOS.  But under DR-DOS of  
the v7.nn range various "DOS executables" have exhibited problems  
at times, yet DR-DOS v7.nn range has a good number of users.

Consequently, at the moment, this 32-bit OpenXP will be classed  
as un-usable by DR-DOS users :(


c.  From a. (above) we can also check/compare such things as  
speeds of process/action with, say, the Windows version of the  
executable.
For an example, currently there's a marked difference in the  
speed of a Nodelist diff update - the Windows version is actually  
much slower.
But that type of "problem" is apparent in a number of Windows  
executables based on a previous DOS program.  For those who've  
known only Windows - they'll possibly never notice it.

Does that help in explaining matters Christian ?

TTFN,

Archie

--- CrossPoint/FreeXP v3.40 RC3 R