Re: Uncle Bob

On Jan 25, 2026, at 7:23 AM, Gad Miron wrote:

Is there any merit in asking IBM to add this to RPG's PSDS?

Just create an idea at the ideas portal - maybe IBM does implement it in the future.

On 25.01.2026, at 15:10 Jim Oberholtzer wrote:

Why? Is it not available with the query? What circumstances occur where you can use the PSDS and the query would not be able to get the status?

Well, first of all, the EXEC SQL needs a lot of time. The query itself and the table function are much slower, than a static marker in the PSDS would be.

Second - SQL is not always an option - OK there still is the API call - but as the job name, user and job number are already available in the PSDS, it only would make sense to make more job info available. E.g . the client IP address would also be great.

Just my 2ct

Regards,
Daniel

Am 25.01.2026 um 15:10 schrieb Jim Oberholtzer <midrangel@xxxxxxxxxxxxxxxxx>:

Why? Is it not available with the query? What circumstances occur where you can use the PSDS and the query would not be able to get the status?

I use the job status all the time but I’ve not run into a situation where it’s not available.


Jim Oberholtzer
Agile Technology Architects

On Jan 25, 2026, at 7:23 AM, Gad Miron <gadmiron@xxxxxxxxx> wrote:
I queried TABLE(QSYS2.GET_JOB_INFO) to get this piece of data
(Batch or interactive).
Is there any merit in asking IBM to add this to RPG's PSDS?
Gad

date: Sun, 25 Jan 2026 08:44:33 +0100
from: Daniel Gross <daniel@xxxxxxxx>
subject: Re: Uncle Bob
IMHO - "hallucinations" - AFAIK there is and never was a "job type" field
in the PSDS.
But it's not uncommon for LLM to "hallucinate" and produce results that
are simply plain wrong.
You then have to "argue" with it to accept the correction.
We still have to accept that - no matter how far advanced those models are
- they are still driven by statistical probability.
And IMHO those hallucination problems will get worse, because they models
are now trained partly with data, that is already generated by other LLMs.
I have about 10 drafts for blog and LinkedIn posts here, that I wanted to
do about Bob since I first seen it at TechXchange - but they all end up not
so positive - so right now I keep them to myself, until I have more
empirical data.
Bob is impressive - some things it really does quite good. But there are
so many cases where it quietly fails, but the code looks so convincing.
Right now, my personal experience says: take everything that ANY LLM/AI is
producing with a good (large) grain of salt!
Kind regards,
Daniel

--
This is the RPG programming on IBM i (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/rpg400-l.
Please contact support@xxxxxxxxxxxxxxxxxxxx for any subscription related questions.

--
This is the RPG programming on IBM i (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/rpg400-l.

Please contact support@xxxxxxxxxxxxxxxxxxxx for any subscription related questions.