Re: [CMS-PIPELINES] PIPEs question

[Jack Woehr]

On 1/25/22 2:31 AM, Rob van der Heij wrote:

This is why "very different" is sometimes
better than "somewhat similar"


True that, friend.

--
Jack Woehr   # Zen is a finger pointing at the moon.
IBM Champion 2021# Some want to see the moon.
http://www.softwoehr.com # Some want to discuss the finger.

Re: [CMS-PIPELINES] PIPEs question

[Rob van der Heij]

On Sun, 23 Jan 2022 at 23:10, Jack Woehr  wrote:

> On 1/23/22 2:55 PM, Rob van der Heij wrote:
> >   The
> > motivation of not having to learn something new is less convincing, I
> think.
>
> Universal axiom, eh, friend?!
>

What I tried to say is that once we translate something like "sed" to CMS
Pipelines context, it wouldn't make CMS Pipelines easier for someone with
another background. They would be troubled by not being able to join lines
by removing \n characters, finding strings that cross record boundaries,
etc. There are certainly use cases for things with regular expressions in
the context of CMS Pipelines, but I would want it for some new function
rather than as alternative syntax for existing functionality.

I have little trouble learning about uniq, tr, sed, awk, grep and such when
I want to do things on my workstation. I don't mind that "tr" does things
that I do with "xlate" in CMS Pipelines; I struggle with when to use \ or
quotes to escape things, and when they mean something else. It would
probably cause me even more trouble when I made "xlate" an alias for "tr"
or so.

Something that frequently bites me is that I'm used to ">" as a built-in
program rather than redirection. It's not unusual for me to wipe out my
file rather than rewrite it :-)  This is why "very different" is sometimes
better than "somewhat similar"

Rob

Re: [CMS-PIPELINES] Will miss CMS/TSO Pipelines

[Rob van der Heij]

On Mon, 24 Jan 2022 at 14:00, René Jansen  wrote:

> I am still baffled by the mores on this list. Please remember that
> imitation is still the sincerest form of flattery.
>

I would argue that's a unique view that's not necessarily shared by large
brand owners, for example. But it will certainly help you dealing with the
loss when someone steals your phone or car. You're free to do as you like,
as long as you don't break the law, upset the lawyers, or whatever. If
you've only used a pair of carpenter pincers to hammer in a screw, then you
may honestly believe that a hammer is compatible with both pincers and
screwdrivers. Look for Dunning-Kruger.

It's been a while since I seriously looked at Ed's work, mainly because he
didn't seem to be aware of concepts in CMS Pipelines that are relevant for
me. If your exploitation of CMS Pipelines is limited to what's there, then
you might be better off to spend some time with various UNIX utilities
(even in cygwin) and basic piping done by the shell. I expect it's easier
to learn about "cut" or "tr" in UNIX than remember the restrictions in
something named after that function in CMS Pipelines.

For serious pipethink on another platform, I probably would look at Python
instead. You will probably find that some of the programming idioms nicely
map to Python functions. It's not multi-stream, but some can be simulated
because objects flow between stages. If we were still doing presentations,
I might be tempted to do one on the analogy between the CMS Pipelines and
Python. Sure, Python isn't REXX, but it shouldn't take you weeks to wrap
your head around it. And you'll find a lot of friends on the Internet.

Don't get me wrong: I would give a lot to have CMS Pipelines on another
platform, but I don't care about something so incompatible that my first
example fails. And I trust John's assessment when he tells me it's not
trivial to do. Also, I'm not sure I have the same use cases there that
would make me as productive as on CMS.

Rob