RE: Re: Modify the IE Browser

> From: Hans Boldt
> 
> I like that phrase "uneasy truce between RPG and the web". Yes, both
> HTTP and RPG are being called into service for things they never
> were designed for. And NOT doing so can result in lost business. Fun
> times we live in, eh?

RPG is doing EXACTLY what it is supposed to do: provide business logic.
HTTP is doing EXACLTY what it is supposed to do: provide transport of
messages from one machine to another.

If you meant "HTML", you might be right that we're stretching its
capabilities a little, but hardly to any great degree.  It's in many
ways like stretching the capabilities of the 5250 data stream to do a
few more things.

And in fact, a browser with a little bit of JavaScript is in many ways
exactly analogous to the 5250 terminals of old.


> Re-architecting applications is something that everyone has to face
> at some time or other. For example, the code base I work with on a
> daily basis is now some 15 years old, and is starting to show its
> age. Many of the design decisions made back then may have made sense
> then, but are now dragging us down.

And many applications don't need to be rewritten.  MRP generation hasn't
changed over the years, nor has basic accounting.  Once you write that
code, it's pretty much in place.

At the same time, most shops DO modify their other business logic.  So
much so that the logic completely reflects their current business
practices, and rewriting it to another language is simply not feasible.
In a majority of real world shops, the biggest asset is not even the
legacy programs, it's the legacy programmers who understand the business
requirements and are able to translate them to program logic.

The best approach in these situations is a staged redeployment from one
architecture to another, not a scratch rewrite.  Replace the UI, then
encapsulate the servers, then remove the middleware.  Simple,
reasonable, doable.  The business logic stays on the host in RPG, and so
your legacy programmers continue to provide the value they always have,
while at the same time you create a place for the UI developers to do
their thing, extending your application to new "paradigms".

Remember, by creating a functioning server-based environment, you enable
not just browsers, but thick clients, wireless protocols, web services,
portals, you name it.  It's just that the browser interface gets you the
biggest bang for the buck and the easiest buy-in from upper management.


> Moving from a green-screen paradigm to a web paradigm is a big shift
> for a lot of reasons. I'm sure a lot of the people here whose
> livelihoods depend on "webifying" tradional OS/400 apps will
> disagree vehemently with me on this point (but so what?), but
> perhaps rewriting apps from scratch is the better choice in most
cases.

Have you ever been involved in a scratch rewrite of a real business
application?  Do you understand the amount of testing that is required?
The fact that no additional business functions can be added during the
period of transition?  The fact that many business rules are not fully
documented?  The fact that you may lose many of your programming staff
who are unable to transition and thus you lose the essential knowledge
of your business practices that these people represent?

I know this stuff intimately, since I was involved in the rewrite of a
portion of BPCS's manufacturing software to a thick client interface.
It took over a man-year just to properly codify the calculations and
design the screens.  And this was for something that was perhaps 15-20%
of the manufacturing functionality.

Business applications are not bits and bytes, they are the embodiment of
business rules that have been shaped and crafted in some cases over
decades.  To state that "rewriting apps from scratch is the better
choice in most cases" flies in the face of experience that most of us
have had in real application environments.  I can count on one hand the
complete conversions from one architecture to another that I know of
that went smoothly, and that includes an RPG II - RPG III conversion I
did when I was a kid.


Finally, enough of the third-hand swipes.  "I'm sure a lot of the people
here whose livelihoods depend on "webifying" tradional OS/400 apps will
disagree vehemently with me on this point (but so what?)".  Okay, who
here DOESN'T think you're referring to me?
 
Of course, if you're NOT referring to me, then you're either referring
to Nathan (though I don't think his livelihood depends on webifying, he
just does it because he thinks it's the right way to go) or to any IBM
WebFacing team members who are here.


Joe