Re: Database I/O Modernization

On Mon, Sep 11, 2017 at 9:59 AM, Nathan Andelin <nandelin@xxxxxxxxx> wrote:

However the API referenced in my original post is not designed for that.
It's merely designed to be a procedural wrapper around RPG I/O op codes.
That is evident by the three (3) parameters passed (Action, Record Buffer,
and Key Buffer).

​And that's the problem. It adds a layer without adding any value. In
fact, it decreases value (by increasing complexity & interference) vs.
just using RPG op-codes in the layer above.

There's two main issues with old monolithic ​RPG.
- business logic built into the UI, so it can't be easily reused, that ends
up copy & pasted into multiple programs.
- dependency on the physical file structure

We could have written "modern" RPG 30 years ago, by simply using a fixed LF
to access the data instead of the PF and by moving logic out to it's own
*PGM. But in particular, the *PGM call would have added overhead.

ILE just removes the overhead, allowing us to use bound calls with little
overhead.

Anything that attempts a one-for-one replacement of RPG RLA opcodes, be it
with ILE procedures or even SQL statements, isn't really modernization from
the DB standpoint.

Sure, tools like Profound allow for UI modernization with a one-for-one
replacement (behind the scenes). But the application itself remains as
crusty and fragile as it's always been...arguably more fragile in fact.

Charles