Win95 Battle Report 
Author Message
 Win95 Battle Report

Good luck on using Win95.

Our experience (with the production release) has been somewhat negative.
To be more explicit, we have a moderate sized image with an OODBMS
(Tensegrity) running under DT ST/VWin (Version 2.0) (For administrative
reasons having nothing to do with this post we have not ported to VSE).

At any rate, after we added the OODBMS, which added more screens etc., we
found that we had broken the ability of Win32S to cope with our problem.
Mysterious exits all the way out of Windows itself were not uncommon. In
desperation we tried Windows NT with great luck. Thus we looked forward
to Win95.

Thursday, we installed the new hype and it seemed to work on our test
bench. Today under load, we started to see mysterious screen behavior:
the most common symptom being "mysterious screen malfunctions" wherein
data would be correctly painted on the screen just in the wrong place (it
always seemed to get painted in the right place as well). The mystery
pixels were in the form of a sprite, i.e. forcing a repaint of the window
removed the offending material. Again, this only happened under severe
screen loading (multiple hidden windows containing complex multiple
views). We are still investigating but our current opinion is that
   Win32s is useless
   Windows NT seems highly functional and stable
   Win95 is somewhere between the two.

I note in passing that Microsoft in some of the non trade press is
hinting that "corporations that can afford it" use NT and not Win 95. At
this point, I would concur. I think that our problems involve the 16 bit
code still resident in the User and GDI DLLs.

BTW, it was curious but our misplaced glyphs always seemed to involve
data from Window Builder Pro's date widget --- not that I am blaming
their code :-)

P.S. we had no trouble running either the development or runtime versions
of ST/V under the production Win95 release. This is a change from the
behavior under the various beta releases.

Any other reports of Win95 behaviorisms would be gratefully appreciated.

--
Charles L. Robinson
Datec Inc.
333 W. Wacker Drive
Suite 1040
Chicago, Ill. 60606-1225
Voice: (312)629-0880
FAX:   (312)629-0883



Wed, 11 Feb 1998 03:00:00 GMT  
 Win95 Battle Report

Quote:
>  We are still investigating but our current opinion is that
>   Win32s is useless
>   Windows NT seems highly functional and stable
>   Win95 is somewhere between the two.

Which Win32s did you try?  In what way was it useless?

Thanks,

Larry
--



Wed, 11 Feb 1998 03:00:00 GMT  
 Win95 Battle Report

Quote:
>Good luck on using Win95.

[problem descriptions deleted]

Quote:
>Any other reports of Win95 behaviorisms would be gratefully appreciated.

If you are doing this for in-house tools you might consider OS/2 Warp
instead.  It is pretty stable and runs your win 3.1 apps.

Terry Raymond



Wed, 11 Feb 1998 03:00:00 GMT  
 
 [ 3 post ] 

 Relevant Pages 

1. 2.1 Report Writer running under Win95

2. Landscape Report to Win95 Generic Print Driver - Save as file

3. Haskell type system battles

4. Delphi V. VO, The battle continues....

5. A diary of a battle: Wild-One, 2.1i and FPGA Express3.3

6. Win95/98 checking and reporting

7. The Battle of the Cultures

8. OpenSSL & Python installed (Half of the battle)

9. popen and spawnv: losing the battle with Win98.

10. Battling Entry Boxes: clue needed

11. NAMELIST vs. bsd f77 - the battle continues

12. Battling bindings

 

 
Powered by phpBB® Forum Software