Re: IBM i "Open I/O" Architecture

Walden H. Leverich wrote:

Mark S. Waterbury wrote:

Why is this idea being presented as only an RPG IV enhancement?
I think it would be far better to provide this capability in
the "base" OS, namely, via a mechanism such as the "override"
commands, so that you could do something like this:

OVRDBF FILE(MYFILE) PGM(MYLIB/MYPGM)
<<SNIP>>

I think this facility should be supported on all of the OVRxxx
commands (OVRDBF, OVRDSPF, OVRPRTF, etc.) -- so that it would be possible to "redirect" any existing file to
another device (or the web).

Problem is, then you'll want it to work for all existing compiled
code, yes? That means you need to intercept the calls that are
already being made, with the existing parameter lists. That's
calls to the QDB* stuff for database and QDM* [sic: QWS & QSF]
stuff for display.
And those are likely some old fragile APIs. Not to mention the
impact that would have on the address resolution caching that
goes on, and I even think that may go through the SEPT. In short,
a cool conceptual idea, but one hell of an engineering task. I
like the new approach better.

Why not that expectation? Such a design would not require maintaining identical parameter lists to the existing programs, because a new architecture could dictate a common parameter list; that, all while leaving all the same old interfaces unchanged, when the alluded override to use the new architecture was not in effect.

As well since the ODP maintains the program to call, the address of the overridden to program could be loaded into the ODP as the call by address.

Since as I suggested in my direct reply to Mark's post, the OS already does almost exactly what he proposes anyway for its standard I/O, extending the concept to a named program to perform the supported methods is really not such a stretch. I agree with Mark that there is little sense to limiting the concept, only to I/O performed from RPG.

Regards, Chuck