Thanks Jonathan for your msg about Som Of The Time, on 05
10-05-1994

We have had a long discussion of SOM on the IBM bulletin board
and really have gone into a lot of its issues. The big one for me
is that those large scale, vertical architectural systems
(persistence, RTTI, object formatting (binary and text), standard
message logging mechanisms, standard exception throwing
mechanisms, ect... cannot be extended over a SOM system composed
of bits and pieces of code written in different languages (because
those languages do not support such mechanisms.) SOM would have to
not define a 'means' for language componentization, but also an
'ends' such that these mechanisms are enforced. Otherwise, it will
never work. If I have built a large system that uses, say, a
standard mechanism for throwing errors as exceptions (an
incredibly powerful system when used with forethought), it makes
no sense for me to then try to use a bunch of modules that have no
such concepts.

For large, multi-programmer system, the standardization of such
mechanisms and styles is paramount for consistent, stable software
development (particularly in our neck of the software woods which
is clinical information systems where being wrong can mean someone
being dead.)

I see SOM as a great way to implement the little productivity and
desktop goodies for OS/2 (as it is now), but it will be a long
time before serious system can be implemented under such a system.

I see something like Taligent, which offers much of the same
capability to piece together components but with the added
advantage of it being a single language (cohesive, coherent)
system, which DOES define many of the needed standards (though
maybe not enough of them.)

___
 X KWQ/2 1.2b X Look Ma, I'm a *Beta Tester*: KWQ Mail/2 for OS/2 v2.0!

--- Maximus/2 2.01wb