hi Justin,
IMHO, you're got that backwards. Those programmers are already
mentally bound to that programming technique. RPG-OA is not the
cause of this, but a consequence.
I think you're right as far as it goes. But I think if it weren't for
RPG-OA (and webfacers, refacing tools, screen scrapers, etc) people
might be forced to update their skills rather than continue with this
paradigm.
On the other hand... I think Henrik is forgetting that there are
millions of already written programs that use the display file paradigm,
and that the amount of work needed to refactor all of those programs
might be prohibitive.
So, in my mind, OA is a great stop-gap solution. It's a band-aid that
helps people get to a new interface quickly ans easily. But it should
not be the ultimate goal.
The ultimate goal should be to write your business logic in a stateless,
services-oriented, MVC manner. That way, it doesn't matter which
front-end you use.
Some people think of RPG-OA as a means of separating the display logic
from the business logic, in order to achieve an "MVC" application -- and
the problem with that is exactly what Henrik said. You aren't writing
your business and display logic separately. You are still writing your
logic as if it's outputting to a proprietary interface that exists
nowhere else, designed originally for 5250 terminals, that are also used
nowhere else.
So that's my opinion, RPG-OA is a great short-term, quick-fix. It's
better than the older technique that involves interpreting the 5250
datastream. But it shouldn't be your end-goal.
Jon Paris constantly makes the point that RPG-OA is useful for other
things besides just user interfaces. He told me this more than 3 years
ago, and I've been thinking about it since then. I've found one or two
cases where this makes sense -- but it's extremely unusual. (And THAT
is why I don't think SPECIAL files ever caught on.)
-SK
midrange.com
