Brian Converse wrote in a message to Kevin Ring:

MB>  KR> I *have* to write Win95 apps, but it would be really great if 
MB>  KR> learning Win95 programming now would make OS/2 programming 
MB>  KR> later much easier.  I would hate to be putting so much effort 
MB>  KR> into skills that are stuck in such a pathetic OS. 
MB> 
MB> My advice is to look at MFC as throwaway effort, worth only what
MB> you are  being paid to use it.  It is not an especially friendly 
MB> tool, it tends to be incompatible with itself from one  version to
MB> the next, and has a number of features which  simply do not work. 

 BC> It may be worth a bit more. There are more employers out
 BC> there who look for people with "MFC" on their resume than
 BC> there are who look for any other class library.

 BC> "Win 95 + VC++ & MFC" is the industry standard. Of course,
 BC> the fact that these employers use such a narrow filter tells
 BC> you reams about what working there will be like.

If your goal to get work writing Windows apps, then MFC is worth knowing. 
But, if your goal is to learn anything useful in connection with
programming OS/2 apps, MFC is a total waste of time.  The Windows and OS/2
PM native APIs have significant similarities because of their common
ancestry, but MFC does not play much of a role in that.  The fact that
Borland was able to get OWL working on both Windows and OS/2 is proof of
what could have been done with MFC, but was not done because Microsoft had
no business interest in doing it.

 BC> RI: careful! That bug's been known to EAT cars on I-95! 

I think there are only about three of us here who understand that line...
 
-- Mike


---