Pick database system and RDBMS
Subhash Jawahrani
I am doing a competative analysis of PICK compared to RDBMS.
Is there a place where i can get some basic info about PICK Database
and 4GL as well? Is Pick a database system or is it a true RDBMS
system. Does it have features like refrential integrity etc? How are
Transaction handled in PICK?
Subhash
John Watson
Subhash Jawahrani <sjaw...@us.oracle.com> asks:
> Organization: Oracle Corporation. Redwood Shores, CA
That's for Pick to know and Oracle to worry about! :-)
Regards,
John Watson
=========================================================================
jwat...@cix.compulink.co.uk | Systems Manager
jo...@nhssne.demon.co.uk | NHS Supplies NE Division
Sequoia OA/CItoh R91/Ultimate | Normanton, West Yorks, WF6 1TL
Henry Eggers
David Rose
>Hello,
> I am doing a competative analysis of PICK compared to RDBMS.
>Is there a place where i can get some basic info about PICK Database
>and 4GL as well? Is Pick a database system or is it a true RDBMS
>system. Does it have features like refrential integrity etc? How are
>Transaction handled in PICK?
no, I'm sorry, we are all to busy developing useful applications for our
users to waste time analysing what makes pick good......
>Subhash
regards,
dave.
------_-_----------------------------------------------------------------------
| o o | David Rose, | .--_|\ |
| ! | Fly By Night Programming, | / \ |
| '''=``` | Australia. | \_.--.*/ |
-------~-----------------------------------------------------------------v-----
Most people I know think that I am crazy (.. Billy Thorpe circa 1973)
Especially the "application programmers" on comp.databases.pick (me circa 1994)
Norman O. Morgan
>I have worked on a few different versions of PICK in the past 8 years
>and have YET to find any "true" transaction processing. I'm talking
>about the kind you get with MS-Access for example, where a BEGINTRANS
>command is issued and all subsequent transactions are performmed in
>memory until a COMMIT or ROLLBACK are issued.
As a Pick newcomer, I was surprised and more than a little disappointed
to find the Pick did not have this transaction rollback recovery. It
does not take an RDBMS to give you this. I have been using it for years
in COBOL with plain old indexed files on both our trusty old Wang VS300
since 1986, and an even more antique Honeywell (remember them?) Level-62
since about 1982. On both systems, all updates cause a before-image
log to be written. When a COMMIT is issued, the log is wiped. If a
ROLLBACK is issued, the before images are restored to the file. In either
case, updated records are locked against further update by other processes
until committed or rolled back. If a system crash or power failure occurs,
a recovery process mass processes all before-image journals to roll all
in-process transactions back to a "clean point" before allowing users
back on the system.
I'll miss that old VS300 in many ways when we are live on the RS6000!
Someone asked me the other day, "When you lose your Wang, do you become
a Unix?"
--
Norman O. Morgan Brake Supply Co., Inc.
nmo...@evansville.net Evansville, Indiana
"Evansville's not Camelot, but it's not Cleveland either."
John Lombardo
: In <3li9v8$2...@dcsun4.us.oracle.com> Subhash Jawahrani
: <sjaw...@us.oracle.com> writes:
: I have worked on a few different versions of PICK in the past 8 years
: and have YET to find any "true" transaction processing. I'm talking
: about the kind you get with MS-Access for example, where a BEGINTRANS
: command is issued and all subsequent transactions are performmed in
: memory until a COMMIT or ROLLBACK are issued.
Transaction processing will be available with the next version of advanced
pick due out mid-year.
A much more important feature, of course, is the ability of the editor to
not print the '.' prompt character between consecutive lines when you press
return without typing a command. That's also in the next version of AP.
Jeff Janner
Norman O. Morgan <nmo...@world.evansville.net> wrote:
>Recently, w...@ix.netcom.com (Don Miller) wrote:
>
>>and have YET to find any "true" transaction processing. I'm talking
>>about the kind you get with MS-Access for example, where a BEGINTRANS
>>command is issued and all subsequent transactions are performmed in
>>memory until a COMMIT or ROLLBACK are issued.
>
>to find the Pick did not have this transaction rollback recovery. It
>does not take an RDBMS to give you this. I have been using it for years
>in COBOL with plain old indexed files on both our trusty old Wang VS300
>since 1986, and an even more antique Honeywell (remember them?) Level-62
>since about 1982. On both systems, all updates cause a before-image
>log to be written. When a COMMIT is issued, the log is wiped. If a
>ROLLBACK is issued, the before images are restored to the file. In either
>case, updated records are locked against further update by other processes
>until committed or rolled back. If a system crash or power failure occurs,
>a recovery process mass processes all before-image journals to roll all
>in-process transactions back to a "clean point" before allowing users
>back on the system.
>
Is this not the way that uniVerse and UniData handle transactions? I know
it is how Prime INFORMATION handled them. If this isn't how it's done,
somebody please explain the difference to me.
--
Jeffrey W. Janner | "There's no good idea that's so good you can't ruin |
j...@abm.austin.tx.us | it with a few well-placed idiots." -- Scott Adams |
Don Miller
>
>Hello,
>
>
> I am doing a competative analysis of PICK compared to RDBMS.
>Is there a place where i can get some basic info about PICK Database
>and 4GL as well? Is Pick a database system or is it a true RDBMS
>system. Does it have features like refrential integrity etc? How are
>Transaction handled in PICK?
>
>
>
>Subhash
>
>
I have worked on a few different versions of PICK in the past 8 years
and have YET to find any "true" transaction processing. I'm talking
about the kind you get with MS-Access for example, where a BEGINTRANS
command is issued and all subsequent transactions are performmed in
memory until a COMMIT or ROLLBACK are issued.
...
The Herrings
Lombardo) writes:
>Don Miller (w...@ix.netcom.com) wrote:
>: In <3li9v8$2...@dcsun4.us.oracle.com> Subhash Jawahrani
>: <sjaw...@us.oracle.com> writes:
[transaction traffic trimmed]
>A much more important feature, of course, is the ability of the editor
>to not print the '.' prompt character between consecutive lines when
>you press return without typing a command. That's also in the next
>version of AP.
A feature which has been part of the Ultimate OS for at least five
years. Guess *that's* what makes the next version of AP 'advanced' ;-)
Sandy
--
(Mr.) Sandy Herring | BRADLEY'S BROMIDE |
Orange County, California | If computers get too powerful, |
always : sher...@ACM.org | we can organize them into committee. |
du jour: pic...@ix.netcom.com | That'll do them in. |
Bob Thomson
>
>Transaction processing will be available with the next version of advanced
>pick due out mid-year.
>
>A much more important feature, of course, is the ability of the editor to
>not print the '.' prompt character between consecutive lines when you press
>return without typing a command. That's also in the next version of AP.
Finally, Pick Systems has learned what a _real_ priority item is ;-}
--
Llucky Llama Games Bob and Debbie Thomson
80-Q North Cabrillo Hwy. #226 (415) 712-0283
Half Moon Bay, CA 94019 llu...@crl.com
Ask about You Rule!, the cool new strategy Play-By-Mail game.
gary valmain
>
> Subhash Jawahrani <sjaw...@us.oracle.com> writes:
>
> >Hello,
> > I am doing a competative analysis of PICK compared to RDBMS.
> >Is there a place where i can get some basic info about PICK Database
> >and 4GL as well?
snip
> no, I'm sorry, we are all to busy developing useful applications for our
> users to waste time analysing what makes pick good......
>
> >Subhash
>
> regards,
> dave.
Dave,
No one has ever accused you of being a wall flower. However, you do
need to learn to open up a bit. You know, tell us what you _really_
thnk about the subject. All of this soft-shoe, dilly-dallying around
will make people think you have no firm opinions.
enjoy
gary valmain
Hal Pawluk
>A feature which has been part of the Ultimate OS for at least five
>years. Guess *that's* what makes the next version of AP 'advanced' ;-)
Sorry, Sandy, but not even close.
The next release, our current 'Ninja' project, is a killer client/server
DBMS. With SQL, ODBC, transaction processing, interoperatibility with
Oracle databases, and a whole lot more. Stay tuned.
Hal Pawluk
Senior Vice President
Marketing & Sales
Pick Systems Inc.
Hal Pawluk
pick due out mid-year.
Not quite: there's no Pick release promised for mid-year, although our
"killer Ninja client/server DBMS" will be well and truly into beta and is
tentatively set for Q4/95 release. It will also include "true transaction
processing" in both Pick Basic and SQL, ODBC, interoperability with Oracle
databases, a Windows GUI (yep), and a whole lot more.
But from now on "no Pick Systems product shall be promised before its time."
Entirely new management, entirely new philosophy.
Jim Idle
> and have YET to find any "true" transaction processing. I'm talking
> about the kind you get with MS-Access for example, where a BEGINTRANS
> command is issued and all subsequent transactions are performmed in
> memory until a COMMIT or ROLLBACK are issued.
You mean... like we do it in jBASE then... smug self satisified and
sickening smirk goes here ;-)
Actually we implemented the same system as in ROS but as the same
chap worte both versions his second attempt is of course much better.
Although I know what you were trying to say by 'in memory' you are
almost but not entirely completely wrong in your statement. The
COMMIT and ROOLBACK are logical concepts, the physical implementation
may be done a number of ways. I will pause here while others jump
in and talk about this. I have very little time left for news these
days and it being Friday night this is seriously cutting into my bar
time. I am already suffering withdrawal symptoms from Spectrum.
Jimi
Jim Idle
>
> >Hello,
> > I am doing a competative analysis of PICK compared to RDBMS.
> >Is there a place where i can get some basic info about PICK Database
Either everyone has assumed that you are here to bash Pick (something
that often seems to happen so you can't blame them), or they don't
think you are serious here. Email me a postal address and I will send
you some glossy marketing brochures about jBASE for he hell of it.
[ Mike (Hannigan), does this count as using the internet as a marketing
vehicle or wot mate? Please flame if so ;-)
]
Jimi
Andrew Shumway
>
> In <3li9v8$2...@dcsun4.us.oracle.com> Subhash Jawahrani
>
> >
> >Hello,
> >
> >
> > I am doing a competative analysis of PICK compared to RDBMS.
> >Is there a place where i can get some basic info about PICK Database
> >system. Does it have features like refrential integrity etc? How are
> >Transaction handled in PICK?
> >
> >
> >
> >Subhash
> >
> >
>
> and have YET to find any "true" transaction processing. I'm talking
> about the kind you get with MS-Access for example, where a BEGINTRANS
> command is issued and all subsequent transactions are performmed in
> memory until a COMMIT or ROLLBACK are issued.
>
>
I'm not sure if it qualifies as 'TRUE' or not but Vmark Universe
has BASIC statements COMMIT, ROLLBACK, END TRANSACTION, BEGIN TRANSACTION
From the description in the online help I think they foot the bill.
They might even be working on a two-phased commit schema for their
distributed locks
Andrew Shumway................................................
David Rose
>Recently, w...@ix.netcom.com (Don Miller) wrote:
>>I have worked on a few different versions of PICK in the past 8 years
>>and have YET to find any "true" transaction processing. I'm talking
>>about the kind you get with MS-Access for example, where a BEGINTRANS
>>command is issued and all subsequent transactions are performmed in
>>memory until a COMMIT or ROLLBACK are issued.
>As a Pick newcomer, I was surprised and more than a little disappointed
>to find the Pick did not have this transaction rollback recovery. It
>does not take an RDBMS to give you this. I have been using it for years
>in COBOL with plain old indexed files on both our trusty old Wang VS300
>since 1986, and an even more antique Honeywell (remember them?) Level-62
the problem is there is no such thing as a free lunch. All this roll back
increases the total system overhead. At least with Pick, as a programmer I
can decide if I want roll-back, and with READUs in BASIC and some sensible
coding techniques, it is rarely required, but I have seen Pick applications
that do their own before-imaging and rollback as required and they have the
freedom to select which file updates attract the overhead.
David Rose
>> Subhash Jawahrani <sjaw...@us.oracle.com> writes:
>>
>> >Hello,
>> > I am doing a competative analysis of PICK compared to RDBMS.
>> >Is there a place where i can get some basic info about PICK Database
>Either everyone has assumed that you are here to bash Pick (something
>that often seems to happen so you can't blame them), or they don't
>think you are serious here. Email me a postal address and I will send
>you some glossy marketing brochures about jBASE for he hell of it.
>[ Mike (Hannigan), does this count as using the internet as a marketing
> vehicle or wot mate? Please flame if so ;-)
>]
>Jimi
maybe I'm a cynic, but when I see an email address like this :-
>> Subhash Jawahrani <sjaw...@us.oracle.com>
^^^^^^
Lights flash all over.
As to to sending him glossies (and telling us you are), doesn't set of my
'marketing-on-the-internet' alarm, I'd only get the pooies if you were to
type it in as your reply.
John Lombardo
: In <3m4ikb$1...@alterdial.UU.NET> jo...@delta1.deltanet.com (John
: Lombardo) writes:
: >Don Miller (w...@ix.netcom.com) wrote:
: >: In <3li9v8$2...@dcsun4.us.oracle.com> Subhash Jawahrani
: >: <sjaw...@us.oracle.com> writes:
: [transaction traffic trimmed]
: >A much more important feature, of course, is the ability of the editor
: >to not print the '.' prompt character between consecutive lines when
: >you press return without typing a command. That's also in the next
: >version of AP.
: A feature which has been part of the Ultimate OS for at least five
: years. Guess *that's* what makes the next version of AP 'advanced' ;-)
: Sandy
: --
: (Mr.) Sandy Herring | BRADLEY'S BROMIDE |
: Orange County, California | If computers get too powerful, |
: always : sher...@ACM.org | we can organize them into committee. |
: du jour: pic...@ix.netcom.com | That'll do them in. |
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Client/Server: (adj.) Computers formed
into a committee.
Henry Eggers
: The Herrings (pic...@ix.netcom.com) wrote:: : >: Subhash Jawahrani <sjaw...@us.oracle.com> writes:
: : [transaction traffic trimmed]
: : >A much more important feature, of course, is the ability of the editor
: : >to not print the '.' prompt character between consecutive lines when
: : >you press return without typing a command. That's also in the next
: : >version of AP.
The question is, can it be turned of or on? :-) If it's 'controlable'
that takes 3 (?) lines of code, else it's a matter of deleting a line.
The more interesting question is, Why did it ever do that?
Regards, hve.
Henry Eggers
: > I have worked on a few different versions of PICK in the past 8 years
: > and have YET to find any "true" transaction processing. I'm talking
: > about the kind you get with MS-Access for example, where a BEGINTRANS
: > command is issued and all subsequent transactions are performmed in
: > memory until a COMMIT or ROLLBACK are issued.
: You mean... like we do it in jBASE then... smug self satisified and
: sickening smirk goes here ;-)
: Actually we implemented the same system as in ROS but as the same
: chap wrote both versions his second attempt is of course much better.
ROS, in this case, being MDIS; Greg's version has undergone
substantial ministrations...
: Although I know what you were trying to say by 'in memory' you are
: almost but not entirely completely wrong in your statement. The
: COMMIT and ROOLBACK are logical concepts, the physical implementation
: may be done a number of ways.
I would observe that 'Pick' machines don't have memory <smirk> or disk
<another smirk>; but, somehow, they retain the before and after versions
of items updated within the transaction, so that the 'before' ends up
on file if 'abort' and the 'after' if 'commit'; and the 'thing'
is concurrency protected from other (correctly behaving) processes while
the transaction is 'open'. And if everything goes south, a well-defined
state can be achieved by some means. The details have varied from time
to time. But the stuff isn't in memory. Except sometimes, but we
don't know when. Primative concept, memory. Do they still have it?
: I will pause here while others jump in and talk about this. I have
: very little time left for news these days
Tell me about it.
: and it being Friday night this is seriously cutting into my bar
: time. I am already suffering withdrawal symptoms from Spectrum.
Always a concern, especially in Portland, with all those aromatics
from the fermentation of hopps and malt in the air...
Regards, hve.