[comp.databases] Pick operating system

paulb@techbook.com (Paul Buder) (09/06/90)

I work for a company which is considering using the Pick Operating System
to keep track of a 500,000 item inventory.   Indexing needs to be done on 
several fields per item.  Additionally each item has to be indexable both
globally, and within its group; there are about 100 groups, and each item
falls into one or more groups.

Does anyone have experience with PICK?  Does it have any major advantages
over UNIX with some form of database management interface?  If not, what
UNIX products would you suggest to build an application? (with C language??)

I am a C programmer in the DOS world.  I haven't had experience with 
extensive db management systems, only simple btree libraries with a dozen 
functions for adding, reading and deleting keys.  Nor have I worked in a
multiuser environment so am not sure how to be more specific, but any help
is appreciated.
-- 
paulb@techbook.COM  ...!{tektronix!nosun,uunet}techbook!paulb
Public Access UNIX at (503) 644-8135 (1200/2400) Voice: +1 503 646-8257
Public Access User --- Not affiliated with TECHbooks

craig@altos86.Altos.COM (Craig Saunders) (09/10/90)

In article <1990Sep6.055815.16617@techbook.com> paulb@techbook.com (Paul Buder) writes:
>I work for a company which is considering using the Pick Operating System...
>

You must work for a smart company.  Pick is an Operating System built around
an efficient database management system.  It's made specifically for business
type applications.

>Does anyone have experience with PICK?  Does it have any major advantages
>over UNIX with some form of database management interface? 

There are many people who have experience with Pick, although I doubt many
of them read the net.  Since my company provides products to both the Pick
and Unix markets, I have the pleasure of interacting with the professionals
in both worlds.  As for your question about the "database management
interface", Pick has one.  I don't believe Unix does.  Unless you add
Informix or Oracle or...  (Of course, some engineers have tried to convince
me that vi is a "database management interface" :-) )

>-- 
>paulb@techbook.COM  ...!{tektronix!nosun,uunet}techbook!paulb

Since getting back into Unix from Pick, I've been monitoring the net for
discussion of Pick.  Since this is the first I've seen, (others say that
it was the point of several flames in one group awhile back) I'll take this
opportunity to make my first posting to the net and ask the question:

Is there anyone else out there with their feet in both worlds (Pick and
Unix) ?  If so, I would like to hear about their experiences.  Areas of
immediate interest are:  using both Pick and Unix in the same processing
environment;  communications between Pick and Unix (Is anyone doing it on
a recurring basis?);  conversions between them;  anyone running Universe
or Pick's new port on the R6000.

{Craig Saunders} {crs@altos.com} {408/432-6200}

<<The opinions expressed above are my own>>

mark@motown.altair.fr (Mark James) (09/12/90)

In article <1990Sep6.055815.16617@techbook.com> paulb@techbook.com (Paul Buder) writes:
>I work for a company which is considering using the Pick Operating System
>to keep track of a 500,000 item inventory.
>[...]
>Does anyone have experience with PICK?  Does it have any major advantages
>over UNIX with some form of database management interface? 

I did a magazine article a couple of years ago comparing multi-user
database management systems for PCs.  The bottom line, between Unix
and Pick at least, was that Unix was more flexible and Pick more
efficient.  BTW, you don't mention your hardware constraints.  Does
this have to work on your company's PCs?  Do they have (or intend to
get) a mini or supermicro?  Do they have a net?  Are they happy with
an ASCII interface, or do they need graphics?  (If they do, Unix is
likely to be better.)  Is response time a major requirement?  Will
they be doing essentially lookup/update things, or complex queries?


In article <4030@altos86.Altos.COM> crs@altos86.UUCP (Craig Saunders) writes:

>There are many people who have experience with Pick, although I doubt many
>of them read the net.

Probably because Pick doesn't cater for TCP/IP (that I know of, anyway).

>As for your question about the "database management
>interface", Pick has one.  I don't believe Unix does.

Pick was *built* around one.  In Unix you have to add a layer, unless
a hierarchical directory structure of flat files suits your needs;
this is why Pick wins in the efficiency contest.  It's also why Unix
is more flexible, since you can choose the DBMS that suits you.  It
also makes for a cost difference, since Unix + Oracle or whatever is
likely to be more expensive than just Pick.

You should also consider whether you want to run other things on the
system than just your database.  Unix has a much larger range of
software available than does Pick.  On a PC, neither system will
coexist very gently with MS-DOS, no matter what the sales types say.

Finally, there are other multi-user operating systems available
besides Pick and Unix (such as Theos, BOS, AMPS, Dataflex, PC-MOS,
Prologue).  These suffer (perhaps unjustly) from a lower profile, and
their database packages tend to gravitate around specific sectors like
clothing manufacturers or insurance, but each has its strengths.

>using both Pick and Unix in the same processing
>environment;  communications between Pick and Unix (Is anyone doing it on
>a recurring basis?);  conversions between them;  anyone running Universe
>or Pick's new port on the R6000.

Products like UniVerse that try to combine Pick data management with
Unix features tend only to marry the efficiency of Unix with the
convenience of Pick -- the worst of both worlds.  UniVerse would be
fine for people who, probably for historical reasons, happen to need
both Pick and Unix anyway, or for developers trying to port their
software from one system to the other.  For folks like Paul, with no
massive investment to protect, it's probably more expensive (in terms
of dollars and performance) than it's worth.

--
=== T. Mark James           ==== All opinions, errors etc are my own.
=== mark@bdblues.altair.fr  ==== "Hardware is that part of a computer
=== +33 (1) 39 63 53 93     ====  system that you can kick."
================================    -- Grace Hopper

stefan@wheaton.UUCP (Stefan Brandle ) (09/12/90)

In article <4030@altos86.Altos.COM> crs@altos86.UUCP (Craig Saunders) writes:
>In article <1990Sep6.055815.16617@techbook.com> paulb@techbook.com (Paul Buder) writes:
>Informix or Oracle or...  (Of course, some engineers have tried to convince
>me that vi is a "database management interface" :-) )

Well, of course it is (lots and lots of smiley faces); I suppose you could
even use it as a database if set up properly (even more smileys).

>Is there anyone else out there with their feet in both worlds (Pick and
>Unix) ?  If so, I would like to hear about their experiences.  Areas of
>immediate interest are:  using both Pick and Unix in the same processing
>environment;  communications between Pick and Unix (Is anyone doing it on
>a recurring basis?);  conversions between them;  anyone running Universe
>or Pick's new port on the R6000.

We're using Pick a la Universe on a DECsystem 5400 as part of our library
automation system (from a company called Dynix).  As far as I can tell,
everything seems to be plodding along pretty well.  However, I don't really get
beyond the Unix-Pick border (I'm a resident on the Unix side), so I'm not
really the person to ask about how well it works.  The general impression I get
from the whistles and wows is that Universe runs at a pretty good clip on the
5400.

-stefan
-- 
---------------------------------------------- MA Bell: (708) 260-4110 ---------
Stefan Brandle                  UUCP: ...!{obdient,uunet!tellab5}!wheaton!stefan
Wheaton College			or	stefan@wheaton.UUCP
Wheaton, IL 60187 		"But I never claimed to be sane!"

prc@erbe.se (Robert Claeson) (09/13/90)

In a recent article mark@motown.altair.fr (Mark James) writes:

>Probably because Pick doesn't cater for TCP/IP (that I know of, anyway).

But Pick-alike products like UniVerse do. UniVerse even implements a
client-server database with UniVerse/Net.

>Products like UniVerse that try to combine Pick data management with
>Unix features tend only to marry the efficiency of Unix with the
>convenience of Pick -- the worst of both worlds.

The last time I checked, no Pick system has won a benchmark over
an Encore Multimax running UNIX System V and UniVerse. This is
not to say that Oracle TPO, Informix Turbo or Ingres on the
same hardware may be even more efficcient.

-- 
Robert Claeson                  |Reasonable mailers: rclaeson@erbe.se
ERBE DATA AB                    |      Dumb mailers: rclaeson%erbe.se@sunet.se
                                |  Perverse mailers: rclaeson%erbe.se@encore.com
These opinions reflect my personal views and not those of my employer (ask him).

nick@aimed.UUCP (Nick Pemberton) (09/14/90)

>Products like UniVerse that try to combine Pick data management with
>Unix features tend only to marry the efficiency of Unix with the
>convenience of Pick -- the worst of both worlds.  UniVerse would be
>fine for people who, probably for historical reasons, happen to need
>both Pick and Unix anyway, or for developers trying to port their
>software from one system to the other.  For folks like Paul, with no
>massive investment to protect, it's probably more expensive (in terms
>of dollars and performance) than it's worth.


Time for a plug.

We use the ADDS MENTOR series of PICK machines, and I wanted to point out
how ADDS has got PICK and UNIX working together. For those unfamiliar with
the ADDS line, their version of the operating system runs on the NCR tower
line of computers.

They have a product out called M/ix, which is essentially NCR Unix (version
3.??.??) with their version of the Pick operating System along side. Both
share a common kernal, and the unix side is responsible for device I/O. The
great thing about this is that the PICK environment is identical to the
Native environment (so much so that even Assembly code can be MLOADed 
directly). Data can be moved between PICK and UNIX, in both directions, under
programmer control. I have seen figures that show a M/ix machine, with no
active unix users is close to 93% as fast as the native PICK machine.

On Monday last I was at a convention where there were two PCs, a M/ix machine
and a native unix machine all tied together via an ethernet (WIN TCP/IP on
the unix and M/ix machine). From the pc, via telnet I was able to log on
to a PICK process on the M/ix machince.  Really slick stuff!

BTW, I am not directly associated with ADDS, I just happen to have a lot
of their hardware!

-- 
Nick Pemberton  		 uucp: !{lsuc, uunet!mnetor}!aimed!nick
AIM, Inc			  bus: (416) 429-1085
Toronto, Ontario, Canada         Home: (416) 690-0647