1 - Yes I didn't look at the samplib
2 - Yes I didn't look at the API
3 - Okay
4 - Okay, whatever you say
5 - Okay

Yes, your observation is wrong. I'm not a unix person and I'm not thinking with 
a "unix mindset", I just asked a question on a topic.

I clarified what I meant. If it helped, fine. If it didn't, if someone was 
interested, they'd have asked to clarify again, not just to prove someone else 
is wrong on the internet.

> For me, someone saying "Parallel run" and "split-up behavior" are referring 
> to thread-safe languages (e.g. GO's goroutines and channels)
That's a heavy assumption man. How is thread safety even in the same area of 
the topic here.
Let me try again... by parallel run, I meant something like A B deployments or 
canary deployments.
In this case, not "deployments" but the concept of going from
Tap A - 100%
Tap A - 100%, Tap B -   0%
Tap A -  50%, Tap B -  50%
Tap A -   0%, Tap B - 100%

> Most Unix programmers don't understand that threads are a small subset of 
> multi-tasking.
Assuming again...
Looking at the rest of this para, I see where you're coming from, wrt assuming 
I'm asking in a unix minded way, because you saw what I was explaining as 
something similar to the unix syslogd "setup". 

But otherwise, it's "funny" isn't it. In z/OS land, everyone has to go to the 
assembler tap. Who knows what % of interfaces are still assembler only. I know 
there's Metal C now but don't think it 100% matches what's possible with 
assembler APIs. So no wonder a unix mindset won't always apply in z/OS as there 
are very few commonalities like that.

It's not that it was insulting, it's the assumption of what the other person is 
trying to say.
Obviously, z/OS is well designed, and has to accommodate decades of small & big 
decisions.
Same for unix/linux. This hobbled together technology is what's powering a 
significant portion of technology.
Possibly even this email server; but since it's IBM-MAIN, it may be on z/VM or 
something.

I was using syslog/operlog interchangeably, my bad; because I meant to only ask 
about the line width aspect of it.



Sent with Proton Mail secure email.

On Saturday, July 26th, 2025 at 22:25, Jon Perryman <[email protected]> 
wrote:

> On Sat, 26 Jul 2025 02:18:06 +0000, kekronbekron [email protected] 
> wrote:
> 
> > I meant to find out the impact of updating IEAMDBLG,
> > and if doing so affects what automation sees;
> > or if IEAMDBLG affects only syslog or only operlog, or both.
> 
> 
> Here are some noteworthy comments
> 
> 1. You didn't look at SAMPLIB(IEAMDBLG) otherwise you would have noted it is 
> a sample that reads OPERLOG.
> 
> 2. You didn't look at the OPERLOG API otherwise you would have realized it 
> "can't affect" anything. It is a READ only API that can't modify OPERLOG in 
> any way.
> 
> 3. OPERLOG is an end destination whereas Automation occurs on the SSI at the 
> message source location.
> 
> 4. I suspect that IEAMDBLG stands for "IEA" = MVS, MDB probably = "Message 
> DeBug" and "LG" = "LoG" (MVS Message DeBug LoG).
> 
> 5. IBM supplies sample IEAMDBLG as one example that the OPERLOG might be 
> useful. IBM lets each site build their own solutions that they find are 
> useful.
> 
> > > > By parallel run, I mean letting the current, split-up behaviour remain 
> > > > the canonical way/source
> > > > How do you know I've mistaken unix threading to multi-tasking?
> > > > How do you know I'm in the unix mindset?
> 
> 
> From your posts, it's obvious. Are my observations wrong?
> 
> You were asked to define "parallel run". Do you think that your definition is 
> correct? Do you think people understood this definition?
> 
> For me, someone saying "Parallel run" and "split-up behavior" are referring 
> to thread-safe languages (e.g. GO's goroutines and channels)
> 
> Most Unix programmers don't understand that threads are a small subset of 
> multi-tasking. For instance, both Unix & z/OS have message handling but in 
> Unix, SYSLOGD collects messages and those messages are redirected to various 
> locations (files, automation, ...). z/OS on the other hand processes messages 
> at the source using the SSI.
> 
> > It's better if you state what you want to say directly than assume 
> > intentions
> 
> 
> I'm sorry you took this as an insult but it was intended as constructive 
> criticism. The sooner you realize z/OS is well designed, the sooner you will 
> think in terms of software design. Unix on the other hand is mostly hobbled 
> together. For instance, consider all the information lost because it doesn't 
> have message IDs.
> 
> > So you're saying leave syslog configs alone, but use operlog as it's richer?
> 
> 
> I'm saying, use what best solves your problem. If SYSLOG meets your needs, 
> then stay with SYSLOG.
> 
> I'm saying SYSLOG will not change. I'm guessing vendors still ask for SYSLOG 
> instead of OPERLOG. I'm guessing that OPERLOG hasn't been life changing for 
> most sites.
> 
> I'm saying that OPERLOG has unrealized potential but it appears most sites 
> can't justify investing in its use. If SYSLOG meets your needs, then stay 
> with SYSLOG.
> 
> Most of the world relies on Unix SYSLOGD which is worse than z/OS SYSLOG. If 
> you need more than z/OS SYSLOG, then you will consider OPERLOG.
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to