A great summation Phil, and accurate. VM (and z/OS) are *comfortable*
environments, because almost everything you can do you can only do one or two
ways, and they are usually pretty darn well documented. In business, this is a
*wonderful* thing. :)
In UNIX, if there are not at least 5 different ways to do something, it is
because nobody anywhere has ever gotten interested in doing it. And the
documentaton is usually either very sloppy, or in a lot of cases, here just is
not any documentation at all. Well, perhaps, there is a usage section in the
code that will display a little help.
The core idea of pipes in Unix was driven by, of all things - economics. To get
the authorization to develop the system at the old Bell Labs, Kerningham,
Ritchie, and company sold UNIX as a text processing system for the copyright or
patent department. (I forget which.) This was on an old DEC PDP system which
very limited memory. Much more limited memory than an IBM mainframe of the day.
To allow multiple users to use the system, they *had* to keep the programs tiny
and sort of stitch them together. Duct tape can fix almost anything I suppose.
In any event, the nroff/troff system, which is a full fledged typesetting
system, also derived from this, and so forth and so on. And since input was on
an ASR-33 teletype machine (imagine wordprocessing on one of those beasties!)
the names of the utilities were kept short because nobody liked typing in those
days. In fact, in those days, some programmers felt it was beneath their
dignity to learn to type, since that was what clerks and secretaries did. I had
a guy who worked for me give me that line as late as 1987!
And underneath all that, the *real* reason was to keep the computer on site -
since all the other ones like the GE GECOS monster the PDP replaced, were
pretty expensive. So to have a computer to develop their ideas on, the
scientists went all out for text processing. AWK came of this as well- with the
initials of the three developers making up the program name. The contention
that it is AWKward to use or AWKward to learn is purely a coincidence. <grin>
And I have a nice bridge to sell too.
VM on the mainframe however, was being driven from different motivations.
Perhaps someone here will share and contrast those reasona and activites for us.
--- Begin Message ---
Paul Raulerson <[EMAIL PROTECTED]> wrote:
>No, I have looked, and CMS Pipelines are nice indeed. But then so are pipes
>under UNIX; indeed, pipes are the very core of UNIX. If you are not annoyed
>by discussing it, I would love to hear your opinions on what is so primitive
>about UNIX. :)
This is an interesting topic: *IX fans are always horrified (at least, at
first) by CMS Pipelines, I think mostly because since Pipes isn't part of the
OS, you have to *gasp* enter a SEPARATE COMMAND to use them. That's obviously
a nit in the scheme of things, but reflects one of the philosophical
differences between the two worlds.
*IX has been described as an environment with a lot of little tools that you
glue (or maybe duct tape) together; CMS is an environment with larger tools
that often don't play together so well. CMS Pipes bridges a lot of that gap,
but (I suspect) since the *IX fans like their paradigm, they see it as a
kludge. Which, in a way, it is.
OTOH, we VMers look at *IX and say "What kind of OS has awk AND sed AND Perl
AND all these obscure little things like 'wc' and 'tr' and 'uniq' and and
and...?" It feels kludgy and awkward (no pun intended) to have so many
overlapping functions.
So maybe this is an "agree to disagree" deal -- the two camps may never really
come together fully.
I look forward to others' thoughts on this topic!
ObAnecdote: 25 or so years ago, back at UofW, we had basic Pipe commands built
into CMS: >, >>, and < at least. They were not a great success; whether this
was due to the lack of the 273 other functions (wc et al.) or due to a
difference in OS philosophy I'm not sure.
...phsiii
--- End Message ---
