Is "Native" Pick(-like) easier than "Open (Pick-like)"?
(Thread taken directly from CDP, with original formatting preserved.)
Hate to break up the techo talk here, but I have a question for people who have
been in Pick for a while (prior to the open systems move). A coworker and I had
a discussion the other day concerning Pick prior to it move to UNIX.

The gist of the discussion was that Pick may have been easier to implement and
maintain before it went to the open system platform. Many of our older
customers did not need System Administrators or an MIS staff. Now that Unidata
and Universe are out, things have gotten more complex.

This is not a knock against the Unix flavors. I realize that for Pick to expand
its market and open up to larger, more sophisticated organizations an open
environment was needed.

Any thoughts on this topic. We found it interesting, and I'm not sure if there
is an answer. Possibly, the system the user buys should fit their needs. Then
there's the case where users want the latest / greatest, rather than the best
and easiest. I've found that its hard to convince a customer the are buying
more than they need! See its an intersting thought.

Michael Prodor



========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: Mike Barnes 
Date: Thu, 25 Jan 1996 17:31:52 +0000

In article <4e88ff$elg@atlas.tncnet.com>, Michael Prodor
 writes
>
>
>Hate to break up the techo talk here,

Go right ahead.

>but I have a question for people who have
>been in Pick for a while (prior to the open systems move).

That's me (1975 vintage).

> A coworker and I had
>a discussion the other day concerning Pick prior to it move to UNIX.
>
>The gist of the discussion was that Pick may have been easier to implement and
>maintain before it went to the open system platform.

No doubt about it, IMO.

>Many of our older
>customers did not need

(or have)

> System Administrators or an MIS staff. Now that Unidata
>and Universe

(and AP over Unix)

>are out, things have gotten more complex.
>
>This is not a knock against the Unix flavors. I realize that for Pick to expand
>its market and open up to larger, more sophisticated organizations an open
>environment was needed.

Sure.  Yesteryear's boxes were very expensive and had very limited
functionality compared with what you get for your money today.

>
>Any thoughts on this topic. We found it interesting, and I'm not sure if there
>is an answer. Possibly, the system the user buys should fit their needs. Then
>there's the case where users want the latest / greatest, rather than the best
>and easiest. I've found that its hard to convince a customer the are buying
>more than they need! See its an intersting thought.

You can get AP (or even R83???) on a PC if you want things the old way.
It's astonishingly cheap compared with its 1970's or 1980's equivalent.
No sysadmin or MIS staff required.  However, it doesn't have what most
of today's users actually want.  [It might have all they *need*, but
that's a different story]

Regards, Mike.

--

Mike Barnes, Owner, Exodus Computer Systems, Stockport, England.
Tel: +44 1663 765734
This week's hot tips for the lottery: 1, 2, 3, 4, 5, 6.

========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: David Ruggiero 
Date: 25 Jan 1996 17:42:16 GMT

mikep@dataworks.com (Michael Prodor) writes:
>The gist of the discussion was that Pick may have been easier to implement and
>maintain before it went to the open system platform. Many of our older
>customers did not need System Administrators or an MIS staff. Now that Unidata
>and Universe are out, things have gotten more complex.

Of course. Pick{ish}-over-Unix systems give you greatly expanded power
and flexibility, but the tradeoff is always increased responsibility,
especially in the system admin realm. As you point out, a system has to be
matched to the user's needs - and abilities.

I've had a hard time over the years trying to make some of my consulting
clients understand that even though the Universe/Unidata/Pick-on-Unix
system some vendor is pitching them as a replacement for their native Pick
box can give them greatly increased capabilities, they'll have to budget
for a 100% (or more) increased system administration costs along with it.
It's no longer a simple matter of having a data entry operator mount a
filesave tape each night, kids. Many of them have a rude shock when they
realize how little they had to worry about all these years with the care
and feeding of their R83ish machine.

David

========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: (Jon Sisk)
Date: Fri, 26 Jan 1996 14:37:39 GMT

mikep@dataworks.com (Michael Prodor) wrote:

>Hate to break up the techo talk here, but I have a question for people who have
>been in Pick for a while (prior to the open systems move). A coworker and I had
>a discussion the other day concerning Pick prior to it move to UNIX.

>The gist of the discussion was that Pick may have been easier to implement and
>maintain before it went to the open system platform. Many of our older
>customers did not need System Administrators or an MIS staff. Now that Unidata
>and Universe are out, things have gotten more complex.

[snip]

One of the Pick Gurus - I believe it may have been Chandru - may have
summed it up best when he said:

  "The _great_ thing about the Pick System is that you can have a
   50-user site with no System Administrator; the _terrible_ thing
   about the Pick System is that you can have a 50-user site with no
   System Administrator".

How true.

From a training perspective, we had it made in the late 70's and 80's
where  "one size fit all". We could have a "mixed" group in _any_ of
our classes. By this, I mean "Generic" Pick, Ultimate and/or
Micro-Douglas-Dyne users. We _did_ have to mention the little
caveats, but they were entirely manageable.

Not so any more. We not only have to segregate the students
by platform ethnicity, but additionally pile on a bunch of "[x"
training, as the "System Administration" issues have been
pulled away from "our" side and moved off to the "[x" side. This
usually requires an additional two weeks to what was formerly a 2.5
week education process.

God bless "[x". It's not just an adventure; it's a career.  


j.
Jon Sisk, JES & Associates, Inc.
   http://www.jes.com/



========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: dessoft@main.com (gary valmain)
Date: 29 Jan 1996 22:13:00 GMT

In article <4eastv$oj5@news2.deltanet.com>, dated Fri, 26 Jan 1996 14:37:39 GMT, Jon
Sisk posits...
>
>mikep@dataworks.com (Michael Prodor) wrote:
>
[snippetted]

>Not so any more. We not only have to segregate the students
>by platform ethnicity, but additionally pile on a bunch of "[x"
>training, as the "System Administration" issues have been
>pulled away from "our" side and moved off to the "[x" side. This
>usually requires an additional two weeks to what was formerly a 2.5
>week education process.
>
>God bless "[x". It's not just an adventure; it's a career.  
>
Such is the price of progress....... It IS progress, isn't it ???

gary valmain
--
_______________________|____independent pick applications programmer____
designed software, inc____|_ladies home companion__wars fought_whiskey_coins
voice_____713.367.8765__|_revolutions started_governments run_bars emptied
fax_______713.367.4316__|_uprisings quelled_burglars alarmed__tigers tamed
email dessoft@main.com__|_country gentleman_orgies organized_coffins_nails
compuserve__72203,2372__|_computers verified___used cars___land___manure
_______________________|_country gentleman___all round really nice guy


========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: Claus Jensen 
Date: 30 Jan 1996 06:39:37 GMT

We have about 60 users plus 16 phantoms.

I hate UNIX but have learned to live it.  I am not technical and can not
program; I can use a current proc and change the english listing.

Two years ago we changed from a sequel to a Motorola Unix running
RealityX.  Here are some good things:

- we have implemented RealityX in such a way that I think that UNIX is
  the ABS section.

- our users login in CAPITAL letters ... just as before

- each users logon sends them to a specific PICK account that can be
  defined by each (UNIX) user login.

- current version of RealityX runs as a raw partition (I think??) under
  unix.

- we have a file save and can do SEL-RESTORES

- to do a 'restore' is simple.

     - logoff all current users and do a SHUTDOWN
     - check that all RealityX processes have been 'killed'
     - at the appropriate promtp type #mdbase something
     - this creates a basic PICK system with about 4 accounts on it
       (one being SYSPROG, POINTER-FILE
     - logout of UNIX
     - logon to SYSMAN (a new account)
     - at TCL do an ACCOUNT-RESTORE ZZ *  (very roughly)
     - this restores the database ... just the the good old times


with this version of PICK, we do not need an administrator; we have fewer
GFE's that we did before

the system is VERY fast

We still have to have programs changed etc.  but we could easily get
along without a MIS type to manage the day to day system.

Education is still required of course ... but RealityX looks just like
the old Microdata ... the spooler is almost identical.

I am still looking for imaging alternatives ... any suggestions???

thanks

Claus Jensen
Vancouver, Canada

ps our company is on the net http://www.overland-freight.com


========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: taj@news (Terry A. Johnson)
Date: 26 Jan 1996 19:04:07 GMT

Michael Prodor (mikep@dataworks.com) wrote:

>The gist of the discussion was that Pick may have been easier to implement and
>maintain before it went to the open system platform. Many of our older
>customers did not need System Administrators or an MIS staff. Now that Unidata
>and Universe are out, things have gotten more complex.

     There are 2 very different issues here - implementing a Pick system
and administering a customer's system.
     My response is based on implementing Microdata Reality and Sequel
and Ultimate "traditional" platforms, implementing Ult+ and supporting
RealityX Unix platforms.  I haven't worked with anything not based on
the original Pick assembly language code.
     On the question of implementing Pick:  I think it is about a wash.
Some aspects are easier and just as many are harder.  Some examples:

     -	Unix takes care of many things that we used to have to do.  A
	lot of time is saved because we no longer have to write a disk
	driver and a disk formatter for every new disk that comes along.
	On the other hand, we do not have the same opportunity to squeeze
	every last bit of performance out of a particular piece of
	hardware.  Unix disk drivers make certain assumptions about how
	programs want to use disks and those assumptions don't always
	match well with what a Pick program might want to do.

     -	Debuggers.  The Pick assembly language debugger had capabilities
	that simply don't exist on the Unix systems.  I miss MLOAD.
	But shared libraries on *some* Unix systems can act like MLOAD.

     -  Tape.  We've generally lost capabilities there.  You would be
	amazed at the hoops we have to jump through to make tape work
	at all.

     -	Maintaining the system.  It's not easier or harder to fix bugs -
	there are just more things to remember.

     -  CPUs.  I remember a discussion with a very good Honeywell
	engineer when I was at Ultimate.  We were developing requirements
	for the follow-on to the dual-15X 7000 platform and trying to
	decide if it made sense to go with commodity CPUs, which were
	about 68030s then or to build one out of custom logic.  He said
	that he could always build a custom processor that would run the
	Pick instruction set faster than any current macro-expansion
	implementation, but that the custom processor would be more
	expensive and the cost difference between the two approaches
	would grow each year.  So, now Pick implementers no longer debug
	CPUs.  That makes things easier, although somewhat less fun.

     The question of administering a system is interesting.  Unix systems
are designed to require a trained system administrator.  Reality and
Ultimate systems were designed to not.
     No two Unix platforms that I have used (HP-UX, AIX, Motorola) are
administered remotely alike.  And each manufacturer changes things from one
release of Unix to the next.  Microdata and Ultimate tried (Ultimate
tried harder) to protect users from a lot of the details of administering
the system.  When Ult+ was introduced, we found that keeping system
administration easy was a battle that we slowly lost.  The company I work
for now, ADP, makes a Herculean effort to make the system that our
clients see easy to administer by hiding the Unix details.  This is a
simplification, but we have a half dozen programmers working on a 6-month
project to, basically, make CREATE-ACCOUNT work better.
     What do we get for all this added complication?  More capabilities,
of course.  Also more opportunities for programmers to create layers of
software that insulate the end user from the dirty details of system
administration.  Maybe those programmers who sold Unix to the world
weren't so bad after all!
     To be fair, there were certainly complications in administering
the old Reality systems.  The skill to get the file restore to skip past
a bad record on tape when it didn't want to was something that people
weren't supposed to need, but it sure came in handy once in a while.
(G213.5, and if that didn't work, G213.7, if I remember right.)
And there were various levels of ability to rebuild the overflow table.
Some people could clear linked overflow, while others could create a
little space out of nothing so a FILE-SAVE could be run.  Yeah, Pick
systems needed gurus.
     Well, this old timer has prated enough - I'd better get back
to CREATE-ACCOUNT.  Sigh...

Terry Johnson

========
Newsgroups: comp.databases.pick
Subject: Re: A Philisophical Pick Question
From: mikep@dataworks.com (Michael Prodor)
Date: 29 Jan 1996 16:10:21 GMT

Alright, excellent feed back. I tend to agree with all of your points. Now, on
the same subject many of the long time Pick people here are not "Computer"
people, but rather "Industry" people. I think this tends to allow (at least our
product) things to be more application oriented rather than language / system
bound. That is, when the system / application was designed it was not done with
Pick Basic in mind, but with the end in mind. Thus it's basics are not platform
specific.

I think that was one of the benefits of working in a Pick environment, compared
to working with a group of straight Computer Science Heads. ;) Again, this may
be the case in many industries, my background is not thast large.

--
Michael Prodor                    "Prepare your Heinie for my Spank Ray"
                        -Space Ghost

Return to CDP FAQ Main Index