Java Packed Decimal

[Benjamin White]

Where is some code to process Z/Series packed decimal?  I know it is not a 
primative data type.  Is there any conversion routines?


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Java Packed Decimal

[Don Higgins]

Benjamin, all

Yes there are Java J2SE compatible open source routines available as part 
of the www.z390.org portable mainframe assembler and emulator project 
which provide standard operations plus conversion between all the 
mainframe data types including: 31 digit packed decimal; floating point 
HFP and BFP short, long, and extended; fixed point 32, 64, and 128 bit 
integers; plus conversion to and from decimal character fields containing 
sign, digits, decimal point, and exponent when required to fit within 45 
character maximum field width.  These formats support numbers in the range 
1E-4932 to 1E+4932 with precision up to 34 decimal digits for floating 
point and 39 decimal digits for 128 bit integers.

The pz390.java emulator uses the Java primitive data types int, long, 
float, and double to support 32 bit integers, 64 bit integers, 32 bit BFP 
and 64 bit BFP floating point wherever possible.  For the data types 
including packed decimal which do not fit within these primitive types, 
the Java classes BigInteger and BigDecimal are used to support arbitrary 
precision numerical values.  The floating point operations are performed 
with extra guard digits and then rounded to the required HFP or BFP 
precision and bit formats.  

With the latest z390 PTF v1.1.01b due out this week, two new macros are 
being included called CTD and CFD to convert between 128 bit integer or 
any of the 6 floating point types and a decimal display character field.  
The default for these macros is to use LINKAGE=SVC for z390 svc 170 and 
171 implemented in Java for speed.  However, LINKAGE=MF for mainframe can 
be specified to call mainframe assembler routines to perform the 
conversions so the macros and routines can be ported to the mainframe as 
required.  Many thanks to Dave Bond with www.Tachyonsoft.com for 
contributing the FPCONVRT floating point mainframe compatible routines.  
These routines have option for z architecture support level with level 5 
indicating use of all the latest instructions.  The routines use many of 
the HLASM more advanced features such as LOCTR and EQU operands 4 and 5 
plus they use new processor features such as 20 bit signed displacements 
so it has been a great compatibility test case for the z390 assembler and 
emulator.  It was also a great test case for compatibility between the 
Java and mainframe floating point support. Two new regression test 
programs TESTFPC1 and TESTFPC2 test conversions covering the entire range 
of floating point values from LB'(MIN)' to LB'(MAX)' using both the SVC 
and MF options.

Don Higgins
[EMAIL PROTECTED]

 from Benjamin While >
A primative data type in a programming language is one that is explicitly 
supportd in the language. A data type that is not made up of other data 
types.  I was referring to the programming language not the hardware 
support

I want to use Java on our Z/Series mainframe.  The Java programs will need 
to read records and databases that have packed decimal data.

Java does not have a packed decimal data type.  I am looking for classes 
or routines that will allow the Java programs to process the paced decimal 
data.

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Java Packed Decimal

[Thompson, Steve (SCI TW)]

I thought I might point out something in this Packed Decimal discussion.

IBM mainframes are not the only computer systems to use PD! Wang VS
systems (ASCII based) also used PD. Burroughs, if memory serves me
correctly, also used PD. I must assume that UNISYS uses PD.

And again, if memory serves me correctly, the S/36, S/38, and AS/400 all
use PD. I'm not sure if Digital Systems (used to be in Walkersville MD)
changed to PD from BCD -- they started out with an ASCII based (bit
sliced) machine that competed against IBM's S/3.

So Packed Decimal is not just an IBM thing.

Later,
Steve Thompson

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Jeffrey D. Smith]

=
-Original Message-
From: "Benjamin White" <[EMAIL PROTECTED]>
Sent: 7/31/2006 7:04 AM
To: "IBM-MAIN@BAMA.UA.EDU" 
Subject: Java Packed Decimal

Where is some code to process Z/Series packed decimal?  I know it is not a 
primative data type.  Is there any conversion routines?
=

I don't know what is your definition of "primitive data type".
It seems to me that when a data type is supported directly
by machine instructions, like Add Packed (AP), Zero and Add
Packed (ZAP), Compare Packed (CP), etc., then it is a primitive
(or "native") data type.

I also don't understand why you are mixing "Java" with
z/Series packed decimal primitive data type. What does
one have to do with the other?


Jeffrey D. Smith
Farsight Systems Corporation
24 BURLINGTON DR
LONGMONT, CO 80501
303-774-9381 direct
303-709-8153 cell
303-484-6170 fax

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Brian Peterson]

Here's what I did.

www.google.com

search criteria:
site:ibm.com java packed decimal

Most interesting hit (based upon 15 seconds of analysis) was a web page 
apparently authored by Mike Cowlishaw (IBM) regarding all of the issues 
surrounding attempting to use floating point arithmetic for commercial data 
processing.

http://www2.hursley.ibm.com/decimal/

First paragraph from the above web page:

"Most computers today support binary floating-point in hardware. While 
suitable for many purposes, binary floating-point arithmetic should not be 
used for financial, commercial, and user-centric applications or web 
services because the decimal data used in these applications cannot be 
represented exactly using binary floating-point. (See the Frequently Asked 
Questions pages for more explanation and examples.)"

Brian

On Mon, 31 Jul 2006 08:04:35 -0500, Benjamin White 
<[EMAIL PROTECTED]> wrote:

>Where is some code to process Z/Series packed decimal?  I know it is not a
>primative data type.  Is there any conversion routines?

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Benjamin White]

A primative data type in a programming language is one that is explicitly 
supportd in the language. A data type that is not made up of other data 
types.  I was referring to the programming language not the hardware support


I want to use Java on our Z/Series mainframe.  The Java programs will need 
to read records and databases that have packed decimal data.


Java does not have a packed decimal data type.  I am looking for classes or 
routines that will allow the Java programs to process the paced decimal 
data.



From: "Jeffrey D. Smith" <[EMAIL PROTECTED]>
Reply-To: IBM Mainframe Discussion List 
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Java Packed Decimal
Date: Mon, 31 Jul 2006 15:31:00 +

=
-Original Message-
From: "Benjamin White" <[EMAIL PROTECTED]>
Sent: 7/31/2006 7:04 AM
To: "IBM-MAIN@BAMA.UA.EDU" 
Subject: Java Packed Decimal

Where is some code to process Z/Series packed decimal?  I know it is not a
primative data type.  Is there any conversion routines?
=

I don't know what is your definition of "primitive data type".
It seems to me that when a data type is supported directly
by machine instructions, like Add Packed (AP), Zero and Add
Packed (ZAP), Compare Packed (CP), etc., then it is a primitive
(or "native") data type.

I also don't understand why you are mixing "Java" with
z/Series packed decimal primitive data type. What does
one have to do with the other?


Jeffrey D. Smith
Farsight Systems Corporation
24 BURLINGTON DR
LONGMONT, CO 80501
303-774-9381 direct
303-709-8153 cell
303-484-6170 fax

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[john gilmore]

The standard device for giving Java access to packed-decimal data is to 
convert these data into double-precision (eight-byte) BFP values.  This is 
always possible, and to the extent that Java is good at any arithmetic, it 
is good at floating-point arithmetic, which it uses very heavily indeed.


John Gilmore
Ashland, MA 01721-1817
USA

_
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Kirk Wolf]

Subsituting BFP for decimals may be standard, but it is usually a bad 
idea (as are many "standard" programming practices) :-(


The standard java package java.math contains a BigDecimal class, which 
is commonly used with JDBC drivers for Decimal columns.

IBM's JDK has an alternative version.

Unfortunately, you have to write your own code to convert a packed 
decimal byte-array field into a java.math.BigDecimal,

but it's not too difficult.

Also, some versions of Websphere Studio Application Developer (aka 
Rational Developer) have (somewhat lousy) support for
converting Cobol copy books into Java classes with field accessors for 
Cobol data types.



Fortunately, there's not too much packed/zoned decimal data on 
mainframes any more :-)


Kirk Wolf
dovetail.com


john gilmore wrote:

The standard device for giving Java access to packed-decimal data is 
to convert these data into double-precision (eight-byte) BFP values. 
This is always possible, and to the extent that Java is good at any 
arithmetic, it is good at floating-point arithmetic, which it uses 
very heavily indeed.


John Gilmore
Ashland, MA 01721-1817
USA

_
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html



--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Thompson, Steve (SCI TW)]

-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On
Behalf Of Kirk Wolf
Sent: Monday, July 31, 2006 2:12 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Java Packed Decimal


Fortunately, there's not too much packed/zoned decimal data on 
mainframes any more :-)



!

Must'a been tongue in cheek.

Later,
Steve Thompson

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Tom Marchant]

On Mon, 31 Jul 2006 14:12:15 -0500, Kirk Wolf <[EMAIL PROTECTED]> wrote:
>
>Fortunately, there's not too much packed/zoned decimal data on
>mainframes any more :-)
>
I don't agree that it is fortunate.  Decimal is the number system
used by the vast majority of humanoid life forms, and IMHO binary
is a poor substitute.  The common decimal number 0.1 converts to
binary as the infinite series
0.000110011001100110011001100110011001100110011001100.

Tom Marchant

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Benjamin White]

Yes, I have been considering extending java.math.BigDecimal.  My first Java 
Z/series programs have converted the packed decimal byte array to Unicode 
for display.


Would extending Big Decimal be the best solution?  Any problems with this?


From: Kirk Wolf <[EMAIL PROTECTED]>
Reply-To: IBM Mainframe Discussion List 
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Java Packed Decimal
Date: Mon, 31 Jul 2006 14:12:15 -0500

Subsituting BFP for decimals may be standard, but it is usually a bad idea 
(as are many "standard" programming practices) :-(


The standard java package java.math contains a BigDecimal class, which is 
commonly used with JDBC drivers for Decimal columns.

IBM's JDK has an alternative version.

Unfortunately, you have to write your own code to convert a packed decimal 
byte-array field into a java.math.BigDecimal,

but it's not too difficult.

Also, some versions of Websphere Studio Application Developer (aka Rational 
Developer) have (somewhat lousy) support for
converting Cobol copy books into Java classes with field accessors for 
Cobol data types.



Fortunately, there's not too much packed/zoned decimal data on mainframes 
any more :-)


Kirk Wolf
dovetail.com


john gilmore wrote:

The standard device for giving Java access to packed-decimal data is to 
convert these data into double-precision (eight-byte) BFP values. This is 
always possible, and to the extent that Java is good at any arithmetic, it 
is good at floating-point arithmetic, which it uses very heavily indeed.


John Gilmore
Ashland, MA 01721-1817
USA

_
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html



--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java packed decimal

[john gilmore]

Benjamin White writes:

|  Yes, I have been considering extending java.math.BigDecimal.  My first 
Java Z/series programs

|  have converted the packed decimal byte array to Unicode for display.

and I am much impressed.  I hope the two single machine instructions

o PKU, PacK Unicode, and

o UNPKU, UnPacK Unicode,

figured at least implicitly in this horrendous effort.

As I noted earlierJava does much of its arithmetic in BFP, some of under the 
covers; and the notion of converting PD values into BFP for use in C/C++ or 
Java environments, while it can certainly be characterized, even 
stigmatized, as a standard practice, is also a good one, the best one at 
least until ASCII DFP becomes available on z/Architecture machines.





John Gilmore
Ashland, MA 01721-1817
USA

_
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java packed decimal

[Jeffrey D. Smith]

-Original Message-
From: "john gilmore" <[EMAIL PROTECTED]>
Sent: 7/31/2006 3:18 PM
To: "IBM-MAIN@BAMA.UA.EDU" 
Subject: Re: Java packed decimal

Benjamin White writes:

|  Yes, I have been considering extending java.math.BigDecimal.  My first 
Java Z/series programs
|  have converted the packed decimal byte array to Unicode for display.

and I am much impressed.  I hope the two single machine instructions

o PKU, PacK Unicode, and

o UNPKU, UnPacK Unicode,

figured at least implicitly in this horrendous effort.

As I noted earlierJava does much of its arithmetic in BFP, some of under the 
covers; and the notion of converting PD values into BFP for use in C/C++ or 
Java environments, while it can certainly be characterized, even 
stigmatized, as a standard practice, is also a good one, the best one at 
least until ASCII DFP becomes available on z/Architecture machines.

John Gilmore


I suggest writing a new class that has one or more factory
methods for creating a BigDecimal from an array of byte. Don't
extend BigDecimal.

2 cents worth. your mileage may vary.


Jeffrey D. Smith
Farsight Systems Corporation
24 BURLINGTON DR
LONGMONT, CO 80501
303-774-9381 direct
303-709-8153 cell
303-484-6170 fax

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Clark Morris]

On 31 Jul 2006 08:31:48 -0700, in bit.listserv.ibm-main you wrote:

>=
>-Original Message-
>From: "Benjamin White" <[EMAIL PROTECTED]>
>Sent: 7/31/2006 7:04 AM
>To: "IBM-MAIN@BAMA.UA.EDU" 
>Subject: Java Packed Decimal
>
>Where is some code to process Z/Series packed decimal?  I know it is not a 
>primative data type.  Is there any conversion routines?
>=
>
>I don't know what is your definition of "primitive data type".
>It seems to me that when a data type is supported directly
>by machine instructions, like Add Packed (AP), Zero and Add
>Packed (ZAP), Compare Packed (CP), etc., then it is a primitive
>(or "native") data type.
>
>I also don't understand why you are mixing "Java" with
>z/Series packed decimal primitive data type. What does
>one have to do with the other?

While I don't know the answer to the original poster's question, bit
is a native data type (more or less) to the z series but until the yet
to be implemented 2002 COBOL standard, COBOL refused to recognize it.
IEEE floating point is another hardware primitive data type that IBM
COBOL short-sightedly refuses to recognize.  I would guess that Java
has no means of describing packed decimal thus not having that data
type as a primitive.

Clark Morris
>
>
>Jeffrey D. Smith
>Farsight Systems Corporation
>24 BURLINGTON DR
>LONGMONT, CO 80501
>303-774-9381 direct
>303-709-8153 cell
>303-484-6170 fax

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Jan MOEYERSONS]

>Java does not have a packed decimal data type.  I am looking for classes or
>routines that will allow the Java programs to process the paced decimal
>data.
>
Google's first hit on packed decimal java:

http://forum.java.sun.com/thread.jspa?threadID=450146&messageID=2043372

Isn't that difficult, is it?

Cheers,

Jantje.

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Bernd Oppolzer]

I had the same problem some time ago with C/370, when DECIMAL was not
supported or was not an option for the installation. So I have some 
routines to handle packed decimal in C. 

I would suggest that you call a JNI (native) routine, written in C or 
another compiler language to handle decimal data. Because IMHO handling the 
decimal data by Java itself would be very slow. 

But: I don't know what overhead the JNI call is, so there will be the need for 
some examination. 

Kind regards

Bernd



Am Montag, 31. Juli 2006 20:07 schrieben Sie:
> A primative data type in a programming language is one that is explicitly
> supportd in the language. A data type that is not made up of other data
> types.  I was referring to the programming language not the hardware
> support
>
> I want to use Java on our Z/Series mainframe.  The Java programs will need
> to read records and databases that have packed decimal data.
>
> Java does not have a packed decimal data type.  I am looking for classes or
> routines that will allow the Java programs to process the paced decimal
> data.
>

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java packed decimal

[Benjamin White]

I found half of what I want in com.dovetail.jzos.ByteUtil.unpackLong().  Now 
I just need a longPacked() method.




From: "Jeffrey D. Smith" <[EMAIL PROTECTED]>
Reply-To: IBM Mainframe Discussion List 
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Java packed decimal
Date: Mon, 31 Jul 2006 23:35:00 +


-Original Message-
From: "john gilmore" <[EMAIL PROTECTED]>
Sent: 7/31/2006 3:18 PM
To: "IBM-MAIN@BAMA.UA.EDU" 
Subject: Re: Java packed decimal

Benjamin White writes:

|  Yes, I have been considering extending java.math.BigDecimal.  My first
Java Z/series programs
|  have converted the packed decimal byte array to Unicode for display.

and I am much impressed.  I hope the two single machine instructions

o PKU, PacK Unicode, and

o UNPKU, UnPacK Unicode,

figured at least implicitly in this horrendous effort.

As I noted earlierJava does much of its arithmetic in BFP, some of under 
the

covers; and the notion of converting PD values into BFP for use in C/C++ or
Java environments, while it can certainly be characterized, even
stigmatized, as a standard practice, is also a good one, the best one at
least until ASCII DFP becomes available on z/Architecture machines.

John Gilmore


I suggest writing a new class that has one or more factory
methods for creating a BigDecimal from an array of byte. Don't
extend BigDecimal.

2 cents worth. your mileage may vary.


Jeffrey D. Smith
Farsight Systems Corporation
24 BURLINGTON DR
LONGMONT, CO 80501
303-774-9381 direct
303-709-8153 cell
303-484-6170 fax

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Benjamin White]

Yes, I also saw that page,  My interest was data type conversions to perform 
calculations, not just display the number.




From: Jan MOEYERSONS <[EMAIL PROTECTED]>
Reply-To: IBM Mainframe Discussion List 
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Java Packed Decimal
Date: Tue, 1 Aug 2006 02:41:11 -0500

>Java does not have a packed decimal data type.  I am looking for classes 
or

>routines that will allow the Java programs to process the paced decimal
>data.
>
Google's first hit on packed decimal java:

http://forum.java.sun.com/thread.jspa?threadID=450146&messageID=2043372

Isn't that difficult, is it?

Cheers,

Jantje.

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Ray Mullins]

The Honeywell (now Bull) Series 6000 also had (has?) a version of PD in its
36-bit word (6 BCD or 4 ASCII characters).

Later,
Ray 

-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Thompson, Steve (SCI TW)
Sent: Tuesday August 01 2006 06:17
To: IBM-MAIN@BAMA.UA.EDU
Subject: Java Packed Decimal

I thought I might point out something in this Packed Decimal discussion.

IBM mainframes are not the only computer systems to use PD! Wang VS systems
(ASCII based) also used PD. Burroughs, if memory serves me correctly, also
used PD. I must assume that UNISYS uses PD.

And again, if memory serves me correctly, the S/36, S/38, and AS/400 all use
PD. I'm not sure if Digital Systems (used to be in Walkersville MD) changed
to PD from BCD -- they started out with an ASCII based (bit
sliced) machine that competed against IBM's S/3.

So Packed Decimal is not just an IBM thing.

Later,
Steve Thompson

--
For IBM-MAIN subscribe / signoff / archive access instructions, send email
to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the
archives at http://bama.ua.edu/archives/ibm-main.html

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Clark Morris]

On 1 Aug 2006 06:17:04 -0700, in bit.listserv.ibm-main you wrote:

>I thought I might point out something in this Packed Decimal discussion.
>
>IBM mainframes are not the only computer systems to use PD! Wang VS
>systems (ASCII based) also used PD. Burroughs, if memory serves me
>correctly, also used PD. I must assume that UNISYS uses PD.
Apparently the Burroughs side of the house no longer has native
decimal arithmetic if I understood Chuck Stevens, Unisys
representative to ANSI X3J4 (COBOL standards) correctly.  
>
>And again, if memory serves me correctly, the S/36, S/38, and AS/400 all
>use PD. I'm not sure if Digital Systems (used to be in Walkersville MD)
>changed to PD from BCD -- they started out with an ASCII based (bit
>sliced) machine that competed against IBM's S/3.
>
>So Packed Decimal is not just an IBM thing.
>
>Later,
>Steve Thompson
>
>--
>For IBM-MAIN subscribe / signoff / archive access instructions,
>send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
>Search the archives at http://bama.ua.edu/archives/ibm-main.html

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Rob Wunderlich]

On Mon, 31 Jul 2006 08:04:35 -0500, Benjamin White 
<[EMAIL PROTECTED]> wrote:

>Where is some code to process Z/Series packed decimal?  I know it is not a
>primative data type.  Is there any conversion routines?
You might want to take a look at the JRIO classes as referenced by this 
post:
http://www.dovetail.com/forum/viewtopic.php?t=108&highlight=packed

-Rob

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Shmuel Metz (Seymour J.)]

In <[EMAIL PROTECTED]>,
on 08/01/2006
   at 09:16 AM, "Thompson, Steve (SCI TW)"
<[EMAIL PROTECTED]> said:

>I must assume that UNISYS uses PD.

Unisys has or had a  successor to the B6500 family, so they must have
supported PD. The 1108 didn't have PD, and I don't know whether the
Unisys successor added it.
 
-- 
 Shmuel (Seymour J.) Metz, SysProg and JOAT
 ISO position; see  
We don't care. We don't have to care, we're Congress.
(S877: The Shut up and Eat Your spam act of 2003)

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Ray Mullins]

Unisys still covers the two different families, so if the Burroughs
equipment had it, then they do support it on that ClearPath line.  (POP for
Unisys, anyone?)

Later,
Ray

-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Shmuel Metz (Seymour J.)
Sent: Wednesday August 02 2006 06:56
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Java Packed Decimal

In <[EMAIL PROTECTED]>,
on 08/01/2006
   at 09:16 AM, "Thompson, Steve (SCI TW)"
<[EMAIL PROTECTED]> said:

>I must assume that UNISYS uses PD.

Unisys has or had a  successor to the B6500 family, so they must have
supported PD. The 1108 didn't have PD, and I don't know whether the Unisys
successor added it.
 

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Clark Morris]

On 2 Aug 2006 09:37:22 -0700, in bit.listserv.ibm-main you wrote:

>Unisys still covers the two different families, so if the Burroughs
>equipment had it, then they do support it on that ClearPath line.  (POP for
>Unisys, anyone?)

As I understood Chuck Stevens who was the Unisys COBOL representative
to ANSI X3J4 and a member of the technical support team for their
compilers, Burroughs large scale B7500 and beyond which I believe
became the A series, stopped supporting full decimal arithmetic and
Mike Cowlishaw claims that only the z series has native decimal
arithmetic and it only has fixed point.  There is no hardware support
anywhere for bankers rounding (round half to nearest even).
>
>Later,
>Ray
>
>-Original Message-
>From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED]
>Of Shmuel Metz (Seymour J.)
>Sent: Wednesday August 02 2006 06:56
>To: IBM-MAIN@BAMA.UA.EDU
>Subject: Re: Java Packed Decimal
>
>In <[EMAIL PROTECTED]>,
>on 08/01/2006
>   at 09:16 AM, "Thompson, Steve (SCI TW)"
><[EMAIL PROTECTED]> said:
>
>>I must assume that UNISYS uses PD.
>
>Unisys has or had a  successor to the B6500 family, so they must have
>supported PD. The 1108 didn't have PD, and I don't know whether the Unisys
>successor added it.
> 
>

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Shmuel Metz (Seymour J.)]

In <[EMAIL PROTECTED]>, on 08/09/2006
   at 04:24 PM, Clark Morris <[EMAIL PROTECTED]> said:

>As I understood Chuck Stevens who was the Unisys COBOL representative
>to ANSI X3J4 and a member of the technical support team for their
>compilers, Burroughs large scale B7500 and beyond which I believe
>became the A series,

The B6500 and B7500 were the original processors in the line that
succeeded the B5000[1] line. As I recall there were B6700, B7700,
B6800 and B7800 processors in the line prior to the merger with UNIVAC
and the renaming of the line.

>stopped supporting full decimal arithmetic

Do you mean that they stopped including the operators in the
processors or only that the compilers stopped generating code for
them?

[1] Despite what you may have read, the B5000, B5500, etc. had
radically different architecture for the B6500 and its friends.
 
-- 
 Shmuel (Seymour J.) Metz, SysProg and JOAT
 ISO position; see  
We don't care. We don't have to care, we're Congress.
(S877: The Shut up and Eat Your spam act of 2003)

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Re: Java Packed Decimal

[Clark F Morris]

On 10 Aug 2006 17:30:49 -0700, in bit.listserv.ibm-main you wrote:

>In <[EMAIL PROTECTED]>, on 08/09/2006
>   at 04:24 PM, Clark Morris <[EMAIL PROTECTED]> said:
>
>>As I understood Chuck Stevens who was the Unisys COBOL representative
>>to ANSI X3J4 and a member of the technical support team for their
>>compilers, Burroughs large scale B7500 and beyond which I believe
>>became the A series,
>
>The B6500 and B7500 were the original processors in the line that
>succeeded the B5000[1] line. As I recall there were B6700, B7700,
>B6800 and B7800 processors in the line prior to the merger with UNIVAC
>and the renaming of the line.
>
>>stopped supporting full decimal arithmetic
>
>Do you mean that they stopped including the operators in the
>processors or only that the compilers stopped generating code for
>them?

Given that Burroughs could require recompilation of programs for
either hardware or software upgrades, I read Chuck's statement as
meaning the processors no longer had decimal arithmetic instructions.
This agrees with what Mike Cowlishaw has on the IBM decimal arithmetic
web site - www2.hursley.ibm.com/decimal (google on IBM decimal
arithmetic if I got it wrong).
>
>[1] Despite what you may have read, the B5000, B5500, etc. had
>radically different architecture for the B6500 and its friends.
> 

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html