CS> This is much more tricky -- storage consistency is hard to
 CS> manage in such an environment.

 MS> objectStore OODB uses this approach too. you have a local object cache,
 MS> and objectStore manages the consistency on all
 MS> machines! (but, the RTFees of objectStore are
 MS> unbeliebable).

Distributed system architecture is a nightmare at this point in time. Maybe
in another decade or two after the academics have exhausted the basic
research topics in this area then distributed systems will be a lot easier
to do.

 CS> Moreover, after years of programming I have built up an
 CS> inherent distrust of code written by other people.  Often
 CS> the interfaces are weird, the code is buggy, and I end up

 MS> :-). Thats exactly my feeling. But, you should read
 MS> the SomObject Toolkit Users Guide. There is simple no
 MS> question on my side after reading it. All
 MS> possibilities are considered. It is well structured.
 MS> really an impressive product and description.

SOM 2.0 is supposed to be included with future OS/2 versions -- it is in
Warp 2, for instance.  But do you know if the SOMObjects toolkit will
become part of the OS/2 toolkit?  Right now it is a separate product, one
which I don't have yet.  Then again, I think it is on one of those DevCon
CD's that I never seem to have the time to thoroughly examine.


--- Maximus/2 2.01wb