Cobol - DFSORT (was: Do you have a Knowledge Officer?

Author

DFSORT (was: Do you have a Knowledge Officer?

William M. Klein

2007-10-06, 6:55 pm

Doc,
Just curious, are you using the "new" (well not that new anymore) facility
for using COBOL "copybooks" and COBOL data-names with DFSORT control cards?
This facility came out well after I actually had to do DFSort work and I was
wondering whether it helps for this type of assignment or not. (I think that
SyncSort added it after DFSort - so both types of mainframe shops should be able
to do it - if others have experience with it and want to comment)

--
Bill Klein
wmklein <at> ix.netcom.com
<docdwarf@panix.com> wrote in message news:fe7j8p$oht$1@reader1.panix.com...
> In article <5moi6gFej77mU1@mid.individual.net>,
> Pete Dashwood <dashwood@removethis.enternet.co.nz> wrote:
>
> [snip]
>
>
> Hmmmm... I've been running into a bit of that lately; I'm on a horrid
> three-part team consisting of company employees, 'stable' consultants
> (folks from a small consulting house looking to expand themselves into the
> next Andersen... errrrr, Accenture or the like) and independent
> consultants... that is, me.
>
> At 9:30 the other morning I got a list of maybe fifty, sixty Oracle
> database column_names and was asked to relate these, by hand, to the
> COBOL-like field names in the documentation for the half-dozen input
> datasets, eg ANN_SAL_OVRTM_AMT is Dataset P8638Y2, OVERTIME-EARNINGS-YTD,
> startpos 866, len 4, PD/COMP-3. I managed my way through this and then
> began to construct DFSORT control statements to create temp datasets from
> the full ones, merge/SPLICE them together and wind up with something that
> could be used to load the database.
>
> So... fifty, sixty fields, half-a-dozen formats, variations in names,
> check for multiple occurrences of the same data in different datasets...
> the usual scut-work. Two and a half hours later I get an email 'we need
> to meet on this'; I assemble my notes, make a few photocopies and into the
> conference-room we go.
>
> The woman who called/leads the meeting is from the consulting-stable...
> and they seem to be looking to get more work from the client in the Old
> Fashioned Way of calling meetings, pointing fingers and generating
> paperwork. She asks what I've done; I pass over a copy of my notes and
> begin explaining... and she cuts me off with 'this will be helpful, do you
> have a copy of it in a word-processing document I can work with?'
>
> I smiled and said 'This is what I have been able to do in the two and
> small change hours since I first laid eyes on this; at that time I
> expected the work to be my notes and not of a caliber immediately grasped
> by any other person at any other time. I can continue coding things up so
> that you'll have your load-file by close of business today...
>
> ... or I can stop doing that and turn to prettying up my working-notes so
> that they can be included in a lovely presentation to be sent off to the
> corner office... in fact, if you give me another *three* hours I can
> probably set them to music. What do you want, the file or a Happy Song?'
>
> And now... and now she no longer invites me to meetings, I have to spend
> my time writing code. It may even be that she thinks this exclusion is a
> punishment to me.
>
> DD

2007-10-10, 9:55 pm

In article <mcSNi.54405$zh6.25635@fe10.news.easynews.com>,
William M. Klein <wmklein@nospam.netcom.com> wrote:
>Doc,
> Just curious, are you using the "new" (well not that new anymore) facility
>for using COBOL "copybooks" and COBOL data-names with DFSORT control cards?

Nope... been too busy cutting the planks to sharpen my saw, if there's a
source you could direct me towatrrds I'd appreciate the URL.

(my Internet access will be spotty for the next w

or so; it may take a
while for me to respond)

DD

William M. Klein

2007-10-11, 3:55 am

Check out:
http://www-304.ibm.com/jct01004c/sy...df/sorttrck.pdf

and look for the section
"Create DFSORT Symbols from COBOL COPYs"

--
Bill Klein
wmklein <at> ix.netcom.com
<docdwarf@panix.com> wrote in message news:fek13b$28m$1@reader1.panix.com...
> In article <mcSNi.54405$zh6.25635@fe10.news.easynews.com>,
> William M. Klein <wmklein@nospam.netcom.com> wrote:
>
> Nope... been too busy cutting the planks to sharpen my saw, if there's a
> source you could direct me towatrrds I'd appreciate the URL.
>
> (my Internet access will be spotty for the next w

or so; it may take a
> while for me to respond)
>
> DD

tleaders...gmail.com

2007-10-11, 6:55 pm

Documentation

Some years (decades) ago I was working on a new system. I spend all my
time coding and testing to develop the system and get it into
production. There was absolutely NO documentation. First I have to
admit that I am a profoundly lazy person, anytime that I am going to
code the same type of program over 2 or 3 times I will write a program
to do the grunt work for me. Here is what I did.

I wrote documentation for each program and then put the documentation
into the source code. Every program in the system has a section in the
Identification Division like this:

*#LOCATION BIN LOCATION LABELS - FOR ZEBRA 2.5x1 AND 4X6
*#LABELS BIN LOCATION LABELS - FOR ZEBRA 2.5x1 AND 4X6
*= LOC410 - BIN LOCATION LABELS - FOR ZEBRA 2.5x1 AND 4X6
*= THIS PROGRAM WILL PRINT BIN LOCATION LABELS FOR THE
*= FOR ZEBRA PRINTER ON 2.5X1 AND 4X6 LABELS.
*=

And in the procedure division like this:

*= "SELECT LABEL SIZE".
*= "ENTER 4 = 4 X 6 INCH LABEL".
*= "ENTER 2 = 2.5 X 1 INCH LABEL".
*=
DISPLAY "ENTER 4 = 4 X 6 INCH LABEL".
DISPLAY "ENTER 2 = 2.5 X 1 INCH LABEL".
ACCEPT WS-LABEL-SIZE.

Lines Coded "*#" are index/cross reference entries. Lines coded "*="
are comments.
I wrote a program read through the source code to extract the
comments, indexes, files, and fields used by each program and store
this information in a series of files.
Then I wrote a documentation print program that used a directory of
the executables as the driving file and printed the documentation.
This way if a program did not have any comments the next time I had
some spare time I could fill in the blanks.

As a new program goes into production the documentation is review by
the production staff. All of their questions are answered by changing
the existing comments, or adding an FAQ section to the source code.

The short form of the documentation containing comments and indexes is
regenerated as a word document every time a source code change is
made. This document is copied to the production staff every night.
The long form of the documentation containing comments, indexes, files
used and file access (read, write, delete. etc) is printed on a
request only.
The full form documentation has never been printed except in test
mode. This contains comments, indexes, files, file access (read,
write, delete. etc), FD data fields used. This information is
available online, but at this level of detail you might as well read
the source code, with one exception. The data field cross reference
will show every program that references a data item. So if I'm going
to make a data base change or research an issue, I know what programs
to review.

So I KNOW that every program in this system is documented.

Tom

2007-10-12, 9:55 pm

In article <GRgPi.100781$TR1.78347@fe06.news.easynews.com>,
William M. Klein <wmklein@nospam.netcom.com> wrote:
>Check out:
>
>http://www-304.ibm.com/jct01004c/sy...df/sorttrck.pdf
>
>and look for the section
> "Create DFSORT Symbols from COBOL COPYs"

Thanks much!

DD

Sponsored Links