On Fri, Jul 15, 2016 at 4:18 PM, Jon Paris <jon.paris@xxxxxxxxxxxxxx> wrote:
"That's mainly an issue of familiarity, though. Listening to IBM i people talk shop is just as arcane-sounding to outsiders, I can assure you.”
Of course it is.
But the point surely is that if OS is to take hold then it is exactly that group of people who have to understand it!! And that is where I think IBM needs some focus.
I can definitely agree with that.
As to tooling yes - it is removing the need for OS knowledge. The term “Unixese” really has nothing to do with the Unix OS per-se but rather the rather strange dialect of English in which they speak.
To an audience that has grown up with commands etc. that follow a pattern that is comprehensible e.g. WRK (Work) OBJ (Objects) PDM (PDM) there is a huge gap to bridge to get to grips with terms such as grep, bourne shell, or better yet bash, awk, git, etc. etc. All of which are mostly meaningless. Add to that that each new thing such as Node.js brings along its own little list of stuff like npm (which to me means New Program Model) and you have a problem.
Well, it's true IBM i commands are more regular. You can often guess
an IBM i command name. You might not always guess right on the first
try, but there are usually only a few possibilities to try, and once
you've exhausted anything that seems intuitive to you, and read
through all the choices that start with intuitive prefixes, chances
are good that the command you want just doesn't exist.
But the midrange platform has always had a centralized style of
development. Each Unix command is a tool written by potentially some
other author. There are tons of shells because tons of people thought
"hey, I can build a better shell, or a shell more to my liking" and
they went ahead and did it, and none was so dominant that it killed
off all the others. It really is a cathedral vs. bazaar kind of thing.
IBM is in a position to say "OK, CPYTOIMPF is the sanctioned name for
the one tool that does what it does; no one can reimplement it or
improve it unless we (IBM) do so."
And commands are just one aspect of the language of a platform. More
broadly speaking, IBM i has things like PTF, LPAR, IPL, and so forth.
You know what these things are purely through familiarity, not through
some pattern in the naming convention. And IBM in general tends to
like their own terminology, even when they are talking about something
that is equivalent or even identical to something folks on other
platforms are talking about. Sometimes they have a good reason (maybe
they invented something on a mainframe ages ago, and newer mainstream
platforms adopted the same thing with a different name, and IBM
chooses to stick with the old name even in their own newer products).
Other times I really get the feeling it's just IBM being willfully
nonconformist or elitist/self-important.
Your point about "the next new thing brings more new terms and
confusion" being a bad thing... it's more a double-edged thing. Along
with the negatives, I think it also encourages creativity and
innovation.
People often bemoan the fact that so little use is made of PASE but never seem to think that perhaps there is a reason. The only people I know who have really tried to make sense of it at all are Scott and Ted (Holt) neither of whom are much involved in this new stuff.
People in the midrange community often bemoan the fact that this
platform has such little share in the market, and it doesn't seem to
be growing. Maybe there is a reason.
Sure, when your installed base is bigger, that's a self-perpetuating
advantage for the incumbent. But look what Steve Jobs did with the Mac
against an overwhelming preponderance of Windows machines. Yeah,
Windows is still more common by a large margin, but Macs are no longer
just for a small niche. And even people who don't own a Mac know what
one is and can recognize one when they see one. If IBM i could become
the Tesla in a sea of Chevys and Toyotas, that would already be quite
something.
John Y.
midrange.com
