Is the (draft) 2008 standard available ?

Per subject.
Is the proposed standard available for inspection anywhere ?

Roger

http://www.cobolportal.com/j4/index.asp?bhcp=1

for a list of documents including the current base, still subject to
the change proposals and further amendment as necessary.

Robert Jones<rjones0@hotmail.com> 05/22/06 1:16 PM >>>
>http://www.cobolportal.com/j4/index.asp?bhcp=1
>
>for a list of documents including the current base, still subject to
>the change proposals and further amendment as necessary.

Specifically, http://www.cobolportal.com/j4/files/std.zip

Interesting bedtime reading.
:-)


---
Frank Swarbrick
Senior Developer/Analyst - Mainframe Applications
FirstBank Data Corporation - Lakewood, CO  USA

Indeed :-)

A couple of things spring to the eye -

Where is the BINARY-SUPER-LONG or whatever
for binary machine format to 128 bits ?
As other numeric formats allow at least 31 digits, this seems
like a glaring omission.
(Incidentally, why 31 digits ?. Logically this should have been 36
following on from 9, 18 ...)

Seems like there is still no facility for interaction with
the environment ie.
ACCEPT FROM ENVIRONMENT or some such equivalent
along with DISPLAY UPON ENVIRONMENT/SET ENVIRONMENT
(or equivalents). Or did I miss something ?

Roger


"Frank Swarbrick" <Frank.Swarbrick@efirstbank.com> schrieb im Newsbeitrag
news:4dep5jF19k7mgU2@individual.net...
> Robert Jones<rjones0@hotmail.com> 05/22/06 1:16 PM >>> 
>
> Specifically, http://www.cobolportal.com/j4/files/std.zip
>
> Interesting bedtime reading.
> :-)
>
>
> ---
> Frank Swarbrick
> Senior Developer/Analyst - Mainframe Applications
> FirstBank Data Corporation - Lakewood, CO  USA

"Roger While" <simrw@sim-basis.de> wrote in message
news:e4uekq$jnf$01$1@news.t-online.com...
\
> Where is the BINARY-SUPER-LONG or whatever
> for binary machine format to 128 bits ?
> As other numeric formats allow at least 31 digits, this seems
> like a glaring omission.

USAGEs BINARY-CHAR, BINARY-SHORT, BINARY-LONG and BINARY-DOUBLE are all
holdovers from '02 COBOL.  Note that neither the '02 standard nor the draft
specifies what these formats are.

As I read the consensus of J4 and WG4 in this area, future formats in which
the format and encoding are entirely up to the implementor -- as is the case
for these, as well as for the older stuff like COMP, DISPLAY and even
BINARY -- are not to be encouraged.

By contrast , USAGE FLOAT-DECIMAL-34 has a precise format defined in the
currently-being-drafted revision to IEEE 754 (therein referred to as
Decimal128).  So does FLOAT-BINARY-34 (Binary128).

> (Incidentally, why 31 digits ?. Logically this should have been 36
> following on from 9, 18 ...)

I don't know where you get nine digits; the first three standards specified
18.  The 2002 standard is what increased it to 31, and the 2008 draft
increases this to 34 in support of the two IEEE 754r floating-point formats,
which as it happens each support 34 decimal digits of precision.   Note that
the draft still only requires implementors to support fixed-point (including
integer) literals of 31 digits or shorter.

The numeric portion of the names associated with the FLOAT-BINARY-n and
FLOAT-DECIMAL-n USAGEs are related to their decimal capacity and, again,
represent formats in IEEE 754r wherein they are defined.

I wasn't particularly happy with this choice of names, but wanted to avoid
characterizations like "short", "medium" and "long" because as is implicit
in your question what was "long" yesterday might be "short" tomorrow.

> Seems like there is still no facility for interaction with
> the environment ie.
> ACCEPT FROM ENVIRONMENT or some such equivalent
> along with DISPLAY UPON ENVIRONMENT/SET ENVIRONMENT
> (or equivalents). Or did I miss something ?

The draft says "The execution of a device format ACCEPT statement causes
information from a device to be transferred to the specific data item, where
the device is a hardware or software device in the operating environment.".
Given the deliberate vagueness in this wording -- that whatever you ACCEPT
from is something that the implementor has defined in the operating
environment that is capable of providing information -- I'm at rather a loss
to understand what it is that you expect of it.

The standard doesn't care whether the source of that information is a CRT, a
PC imitating a CRT, another program (whether simulating a hardware device or
not), an ASR33 teletype, a Friden Flexowriter,  or a rutabaga.  What part of
the "ENVIRONMENT" are you expecting to receive data from, send data to, or
"set"?

-Chuck Stevens

"Chuck Stevens" <charles.stevens@unisys.com> wrote in message
news:e4vf9v$2kvc$1@si05.rsvl.unisys.com...
> "Roger While" <simrw@sim-basis.de> wrote in message
> news:e4uekq$jnf$01$1@news.t-online.com...
> \
<snip> 
>
> The draft says "The execution of a device format ACCEPT statement causes
> information from a device to be transferred to the specific data item, whe
re
> the device is a hardware or software device in the operating environment."
.
> Given the deliberate vagueness in this wording -- that whatever you ACCEPT
> from is something that the implementor has defined in the operating
> environment that is capable of providing information -- I'm at rather a lo
ss
> to understand what it is that you expect of it.
>
<snip>
>    -Chuck Stevens

Roger,
Chuck doesn't come from an X/Open -type COBOL environment, so he didn't
understand your question. Your real question was "how do you get and set
run-time environment variables".  (The X/Open syntax has always been
"conforming" - but only portable to SOME implementations and environments).

I know that there have been discussions about this at both J4 and WG4 (i.e.
adding a "standard syntax" for doing this) - along with a function (or
something) to get the number of arguments passed to a program/function.   I
agree that I don't see this in the draft '08 Standard.  Once he understands 
the
question, Chuck might be able to tell us whether this is still on the
"candidates" list - or has been rejected (and if so why).



--
Bill Klein
wmklein <at> ix.netcom.com

"Roger While" <simrw@sim-basis.de> wrote in message
news:e4uekq$jnf$01$1@news.t-online.com...
[snip]
> Seems like there is still no facility for interaction with
> the environment ie.
> ACCEPT FROM ENVIRONMENT or some such equivalent
> along with DISPLAY UPON ENVIRONMENT/SET ENVIRONMENT
> (or equivalents). Or did I miss something ?

Mr While, what you seemed to have missed is that the
draft COBOL standard was not written specifically for
POSIX; but the draft COBOL standard does permit an
implementor to interface to POSIX.

As I understand the current draft COBOL standard and
with reference to WD 1.6:

Page 215, 12.2.6.1 SPECIAL-NAMES paragraph, GF,
"device-name-1 IS mnemonic-name-3".

Page 217, 12.2.6.2 SPECIAL-NAMES paragraph, SR 7,
"Mnemonic-name-3 may be specified only in the ACCEPT
and DISPLAY statements. The implementor may specify
additional restrictions on the use of mnemonic-names
that reference specific device-names."

Page 217, 12.2.6.2 SPECIAL-NAMES paragraph,
SR 8, "The implementor shall specify the names that are
available for switch-name-1, feature-name-1, and
device-name-1."

The implementor specifies device-names and how to use
them. The programmer associates the device-name
(system-name) to a mnemonic-name (user-defined word)
in the SPECIAL-NAMES paragraph.

Thus an implementor could define the following:

POSIX-ARGUMENT-COUNT
POSIX-ARGUMENT-NUMBER
POSIX-ARGUMENT-VALUE
POSIX-ENVIRONMENT-NAME
POSIX-ENVIRONMENT-VALUE
POSIX-SYSERR

The programmer may then use these device-names in the
SPECIAL-NAMES paragraph to access the "device".

SPECIAL-NAMES.
POSIX-ARGUMENT-COUNT IS ARGUMENT-COUNT
POSIX-ARGUMENT-NUMBER IS ARGUMENT-NUMBER
POSIX ARGUMENT-VALUE IS ARGUMENT-VALUE
POSIX-ENVIRONMENT-NAME IS ENVIRONMENT-NAME
POSIX-ENVIRONMENT-VALUE IS ENVIRONMENT-VALUE
POSIX-SYSERR IS SYSERR.

Then in the PROCEDURE DIVISION.

ACCEPT MY-ARGUMENT-COUNT FROM ARGUMENT-COUNT.
DISPLAY ZERO UPON ARGUMENT-NUMBER.
ACCEPT THIS-PROGRAM-NAME FROM ARGUMENT-VALUE.
DISPLAY "data_directory" UPON ENVIRONMENT-NAME.
ACCEPT DATA-DIRECTORY FROM ENVIRONMENT-VALUE.
DISPLAY "Program Error: " PROGRAM-ERROR-NUMBER
UPON SYSERR.

Having used only some of the X/Open (POSIX) features
and only rarely, this is my best guess as to what is may look
like in a conforming implementation.