* copyright (C) 1984-2019 merrill consultants dallas texas usa

May 18, 1993 are tested for the OUTPUT DO group:

IF VSRNTRKR GT 0 OR VSRNTRKW GT 0 OR VSRNDSV GT 0 OR VSRNBYTR GT 0

OR VSRNBYTW GT 0 OR VSRNDSF GT 0 OR VSRNVOL GT 0 OR VSRNSYS GT 0

OR VSRTAGE GT 0 OR VSRTTINQ GT 0 OR VSRTTINP GT 0 OR VSRTTWV GT 0

OR VSRTCPU GT 0 THEN DO;

Thanks to Neil Campbell, Inland Revenue, ENGLAND.

Change 11.076 Lines 001100 thru 001700, the OUT= data set name DCOLMIGS

DAILYDSN was repeated, but it should be the same data set name as

May 18, 1993 the DATA= parameter on the same line.

Thanks to Ian Jones, The Equitable Life Assurance Society, ENGLAND.

Change 11.075 Lines 006300 and 006600 both contain DATA=TYPEVVDS, but

ANALVVDS that should have been DATA=_LTYVVDS in both lines, to be

May 18, 1993 consistent with the new _L macro naming conventions.

Thanks to Ian Jones, The Equitable Life Assurance Society, ENGLAND.

Change 11.074 The new member CHANGESS now contains ALL changes from ALL

CHANGESS MXG Versions, which should make online browse a little

May 15, 1993 easier. Current Changes are still in member CHANGES.
Change 11.073 Boole's IMF MSC Arrival Time was mis-documented. The

VMACCIMS three lines reading:

May 15, 1993

@216+OFFIMS TH PK1.

@;

SEC=SEC+(TH/100);

must be changed to read:

@216+OFFIMS TH ?? PD1.

@;

SEC=SEC+(TH/10);

Thanks to Silvio Orsini, Banca d'Italia, ITALY.

Thanks to Vladimiro Viseno, Assicurazioni Generali, ITALY.

Change 11.072 The code to set MULTRANS='Y' set NMSGPROC=1 only for the

TYPEIMSA first transaction in the group. The code was relocated

May 15, 1993 and a retained flag used to correctly reset NMSGPROC.

Thanks to Lonnie T. Rimmer, Philip Morris, USA.

Change 11.071 Candle's ITRF (a part of Omegamon/IMS Version 110) data

VMACITRF records EVENTIME was incorrect.

May 15, 1993
Change 11.070 STOPX37 Release 3.5 data records do not agree with the

VMACX37 DSECTS provided by EMPACT. Undocumented bytes caused

May 15, 1993 INVALID value for JINITIME. Many lines were changed.

Thanks to Dave Redinger, General Electric Plastics, USA.

Change 11.069 Continued enhancement of RMF look-a-like reports from MXG

ANALRMFR corrected several syntax errors and added additional

FORMATS reports, and formerly temporary formats are now built by

May 19, 1993 new PICTUREs in member FORMATS.

Thanks to Norbert Korsche, OMV-AG, AUSTRIA.
=Changes thru Change 11.068 were included in MXG Early Prerelease 11.01=

===================that was dated May 15, 1993==========================

Change 11.068 Support for Corporate TIE user SMF record creates six new

EXTYTIE1- MXG datasets, reporting file transfer activity between

EXTYTIE6 PCs and Mainframes with this product:

IMACTIE TYPETIE1 - TIE Initialization/Termination

TYPETIE TYPETIE2 - TIE Logon/Logoff

VMACTIE TYPETIE3 - TIE Virtual Disk OPEN/CLOSE/MOUNT/DISMOUNT

May 11, 1993 TYPETIE4 - TIE File Transfer Send/Receive Start/End

TYPETIE5 - TIE QIMPORT/QEXPORT Start/End

TYPETIE6 - TIE VDIMPORT/VDEXPORT Start/End

Thanks to K. M. Torsvik, Postverkets Datasentral, NORWAY.

Change 11.067 This "split" architecture for BUILDPDB/BUILDPD3 has been

BUILD002 further enhanced. By splitting BUILDPDB into separate

BUILD004 pieces, parallel steps could be used to reduce elapsed

BUIL3001 run time, or some pieces could be skipped, etc. This is

BUIL3005 still in testing and the design may still change.

BUILD006 The CICS sorting that was in BUILD002 was moved to

May 11, 1993 BUILD004, the colon in BUILD003 was changed to semicolon,

new JES3 members BUIL3001 and BUIL3005 were created, and

BUILD006 was deleted. The revised architecture:

To replace BUILDPDB (for JES2):

BUILD001 - Reads SMF and creates all WORK data sets,

CICSTRAN.CICSTRAN, DB2ACCT.DB2ACCT and the

PDB.DB2STATn data sets.

BUILD002 - SORT some WORK datasets into PDB library.

BUILD003 - SORT RMF datasets into PDB and then build

PDB.RMFINTRV dataset.

BUILD004 - Combine CICS Statistics datasets and create

four PDB.CICxxxRV datasets.

BUILD005 - Sort TYPE30xx, TYPE6, TYPE26J2, interleave

with SPIN library, create PDB.JOBS,PDB.STEPS

PDB.JOBS, update SPIN library, copy SPIN

datasets into PDB (for backup), and create

PDB.SPUNJOBS.
To replace BUILDPD3 (for JES3):

BUIL3001 - Same function as BUILD001, for JES3.

BUILD002 - Same as JES2

BUILD003 - Same as JES2

BUILD004 - Same as JES2

BUIL3005 - Same function as BUILD005, for JES3, plus

sorts TYPE25 dataset.
One immediate use of this "split" architecture is for

VM/370 sites running BUILDPDB that do not have CICS data;

the virtual size of BUILDPDB can be reduced by removing

CICS processing. You need only to create a VMAC110 in

your USERID.SOURCLIB member with this line:

MACRO _VAR110 % MACRO _CDE110 %

and then instead of %INCLUDE SOURCLIB(BUILDPDB) use

%INCLUDE SOURCLIB(BUILD001,BUILD002,BUILD003,BUILD005);

It is also possible to remove DB2 processing, but as only

three datasets are created, the saving is much less. To

remove DB2, you would create a VMACDB2 in USERID.SOURCLIB

MACRO _VARDB2 % MACRO _CDEDB2 %

and also create a member DIFFDB2 with a blank line. Then

the preceding %INCLUDE would produce the minimum virtual

storage size by eliminating both DB2 and CICS processing.

Under MVS, the SMF processing (BUILD001) needed virtual

MEMSIZE of 17301K. With no CICS only 11397K was required,

and with neither CICS nor DB2 only 9973K was required.

Thanks to Neil Campbell, Inland Revenue, ENGLAND.
Change 11.066 Variables RDMT, RDTBK1-RDTBK4, and RNVSOV were not in the

VMACDMON KEEP= list, but have now been added for ASTEC datasets

May 11, 1993 DMONVOL, DMONJOBS, and DMONDSN. RDMT was formatted.

Thanks to Rodney Marwick, Vereinte Versicherungen, GERMANY.

Change 11.065 OPC/A data records may be lost; line 836, which reads:

VMACOPC IF _OPCVER AND MTDTYPE EQ 9 THEN INPUT +2 @;

May 11, 1993 should be deleted. See also later Change 11.122.

Thanks to Rodney Marwick, Vereinte Versicherungen, GERMANY.

Change 11.064 Migrated datasets with ORGEXPDT=9999999 were not expected

VMACDCOL in MXG's calculation of EXPDT='01JAN2099' for never

May 11, 1993 expiring datasets, but now that test is included.

Thanks to Alan M. Sherkow, I/S Management Strategies Ltd., USA.

Change 11.063 DB2 format MGDB2TY did not decode the '0A'X value. Also,

FORMATS values of zero for QWHCATYP occur in data records, but

May 10, 1993 are not documented by IBM; since all 0 occurrences have a

value of UTILITY for both QMDACNAM and QWHCCN, I assumed

that value, and added two lines to the format:

0='0:UTILITY'

0AX='AX:IMS TRANSACTION BMP'

Thanks to Rod Brady, Zurich Insurance, ENGLAND.

Change 11.062 CONFIGVM is a configuration file for MXG under CMS SAS.

CONFIGVM It is a modification to the CONFIG07/CONFIG08 MVS options

May 10, 1993 and has been found to support execution of BUILDPDB in a

32MB virtual machine.

Thanks to Tom Maynard, Nalco Chemical Company, USA.
Change 11.061 Candle's IMS Transaction Reporting Facility has now been

VMACITRF validated with data, and some revisions were required.

May 10, 1993 Five data sets are created:

ITRFMSG - One observation for each IMS transaction;

a transaction is defined by ITRF as a

message received and response sent, or a

program scheduled, processed, and ended.

ITRFMSGO - One observation for each message sent to an

alternate TP PCB by another transaction.

ITRFDATA - One observation for each database call (DLI

or FP)

ITRFSUM - One observation per database called per

transaction.

ITRFDB2 - One summary record per DB2 transaction.

Note that variables QWACNID and QWHCCV are

created in ITRFDB2 so that you can match the

IMS transaction in ITRFDB2 with its DB2ACCT

counterpart transaction.

Change 11.060 For jobs that initiate instantaneously in a multi-system

VMAC30 environment in which system clocks are not synchronized,

May 10, 1993 elapsed duration variables JELAPSTM,SELAPSTM,DSENQTM,

ALOCTM,EXECTM may be wrong (large positive or negative),

and the date of INITTIME,ALOCTIME,LOADTIME,JINITIME may

be one day later than reality! Change 10.213 (MXG 10.3+)

attempted to detect and circumvent an un-fixed IBM APAR,

(an obscure problem after an A78 ABEND), but my cure was

worse than the disease if REND (reader end time) is later

than INITTIME. The correction is to remove Change 10.213:

Find "10.213" and remove the four lines of that DO group.

(For reports with JELAPSTM from PDBs already built, you

can recalculate JELAPSTM=JENDTIME-JSTRTIME;)

Thanks to Shaheen Pervaiz, Trans Union Corporation, USA.

Change 11.059 Support for ZARA, The Tape Media Manager from Altai. This

EXZARDSN product is a replacement for CA-1 (TMS), and this support

EXZARVOL (originally written by Miller Dixon) processes their tape

IMACZARA catalog records to create two MXG datasets useful for the

TYPEZARA chargeback of tape data sets and for capacity planning of

VMACZARA tape volumes:

May 7, 1993 ZARADSN - One obs per dataset on tape, with attributes,

TAPEFEET (estimated), and volume account code

from the corresponding VOL record.

ZARAVOL - One obs per tape volume, with summed TAPEFEET

for all NRDSETS on the volume, and with the

maximum EXPDATE, CRDATE, and last use date of

any data set on this volume.

Member TYPEZARA has example JCL for the ZARA utility that

is executed to create the data records processed by this

support.

Thanks to Miller Dixon, Continuum, USA.
Change 11.058 The MULTIDD='Y' flag was not set before the IMAC30DD exit

VMAC30 was taken, causing MULTIDD to always be blank in dataset

May 7, 1993 TYPE30_D (also PDB.DDSTATS). The block of code which set

MULTIDD was moved to immediately before the IF OFFEXCP...

block of code.

Thanks to Diane Eppestine, Southwestern Bell, USA.

Change 11.057 DCOLLECT SMSDATA (SMS constructs) records cause STOPOVER.

VMACDCOL As indicated in Change 10.309, no test data was available

May 6, 1993 for these new records. Now that I have test data records

I have corrected for my errors (as well as some of IBM's

mis-documentation) and believe the support is now valid

for seven subtypes (DC,SC,MC,BC,SG,VL,AG). The two OAM

subtypes (DR,LB) apply only for optical drives and were

not available for testing, but should also be valid now.

Thanks to Chris Coldiron, Depository Trust Company, USA.
Change 11.056 Support for SYNCSORT Release 3.5 added 27 new variables,

VMACSYNC including expanded storage usage by the sort, and EXCP

May 6, 1993 counts were relocated to the end of the record.

Thanks to Sam Sheinfeld, Kraft General Foods, USA.

Change 11.055 This "PRINT ALL" utility still used POINT= statement, but

VMXGPRAL that fails with indexed datasets in Version 6, so the

May 6, 1993 logic was rewritten to not use POINT=syntax, just in case

you wanted to print an indexed dataset.

Change 11.054 TYPE80A can still fail with INPUT STATEMENT EXCEEDED.

VMAC80A Change 11.017 fixed only one case, but each INPUT with

Apr 30, 1993 "$VARYINGzz. yyyyyyy" syntax must be protected in case

yyyyyyy is greater than zz. (The zz value defines the

length of the variable; IBM docs many of the RACF fields

as 1-255, but MXG, in trying to minimize the stored size

of the TYPE80A dataset, picked more reasonable zz values.

For those fields that exceed the MXG zz choice, an error

occurs). The solution is to insert these two lines:

SKIP=yyyyyyyy-zz;

IF SKIP GT 0 THEN INPUT +SKIP @;

after the @; at the end of each INPUT statement that uses

the $VARYING informat. In addition to this protection,

all of the fields with $VARYING44. or $VARYING32. were

changed to $VARYING64. to minimize data loss in those

cases in which data length was more than MXG's default.

Thanks to Joseph J. Faska, Depository Trust, USA
Change 11.053 NOTE: INVALID NUMERIC DATA, 'AVERAGE*ACTIVE-ONLY...'

VMACVMXA occurs because at line 094370 there is a semi-colon after

Apr 28, 1993 the label for variable AVGCONMS, but that is not the last

variable in the label statement. Remove the extra ';'.

Thanks to Peter Smith, Sema Group FM, ENGLAND.
Change 11.052 Specifying an undefined INTERVAL= value did not warn you

VMXGDUR that MXG did not recognize your request, and fell thru.

Apr 28, 1993 Now, a warning message is printed.

Thanks to Ed Lau, Lucky Stores, USA.

Change 11.051 The text of Change 11.001 was revised (the 8-lines to be

VMAC37 replaced was not exact, but there was no error in the 12

Apr 28, 1993 replacement lines). Additionally, the INPUT of OFFDETT

must be PIB4. instead of PIB2.

Thanks to Rod Fernandes, Albert Heijn B.V., HOLLAND.

Thanks to Irmgard Polifka, SAS Institute, GERMANY.

Change 11.050 DB2ACCT variable NETSNAME was incorrectly padded with

VMACDB2H blanks, and did not match CICSTRAN variable NETSNAME,

Apr 28, 1993 which is padded with nulls. This prevented merging DB2

and CICS transactions. Find "SUBSTR(NETSNAME...", and

insert just before that line this statement:

NETSNAME='0000000000000000000000000000000000000000'X;

(that's 20 bytes or forty zeros). Then, thirteen lines

later, replace ELSE IF 1 THEN DO;

with ELSE IF LOCBLANK EQ 1 THEN DO;

Now, variables NETSNAME and UOWTIME can be used to match

DB2ACCT observations with their CICSTRAN counterparts.

Thanks to Ron Thein, Mortgage Guaranty Insurance Corp., USA.

Change 11.049 Support for HMF, Host Monitoring Facility Product for

EXTYHMF0-9 Channel-Extension Hardware Monitoring. HMF creates SMF

IMACHMF records with 11 subtypes; this code supports subtypes 0

TYPEHMF and 6 thru 10. This is from Network Systems, known as

VMACHMF DXE's and sometimes called HyperChannel

Apr 27, 1993

Thanks to Bob Hursch, Lockheed Information Technology, USA.
Change 11.048 "ERROR 455-185 dataset TYPE30OM not specified" because

ANALDSET the override for that new dataset was not added to this

Apr 27, 1993 analysis member. Add after the last /* OVERRIDE ...

XY ADD NAME=EXTY30OM

/* OVERRIDE AND OUTPUT NO TYPE30OM RECORDS */

Thanks to Matt Klett, Kohler Company, USA.

Change 11.047 VM/ESA "UNEXPECTED/INVALID CONTROL RECORD" message occurs

VMACVMXA and no observations are created, because MXG's validity

Apr 27, 1993 check tested four bytes for 00008709x, but the first two

bytes may now be non zero. Change the test:

IF IPARMLF1 NE 00008709X THEN DO;

to read:

IF MOD(IPARMLF1,10000x) NE 8709X THEN DO;

Thanks to Walter H. Clark, C.U. Insurance, USA

Change 11.046 Variable SYSTEM was specified twice in the KEEP= list of

VMACTMVS datasets TMVSIV, TMVSJI and TMVSNQ. This caused a SAS

Apr 27, 1993 warning message, but no real impact.

Thanks to Rod Fernandes, Albert Heijn B.V., HOLLAND.

Change 11.045 Variable ASISDCDY should have been spelled ASISDCDV in

ZRBRPT1 members ZRBRPT1 and ZRBRPT2.

ZRBRPT2

Apr 27, 1993

Thanks to Marcel Frechette, Montreal Trust Technology, CANADA.
Change 11.044 Variable QTXANPL exists only in DB2ACCT, and should not

ADOCDB2 have been in the KEEP= list for DB2STAT1. It caused no

DIFFDB2 error, only confusion because it was missing in DB2STAT1.

VMACDB2 It was removed from KEEP= list, and from documentation.

Apr 27, 1993

Thanks to Ellen Ulrich, Texas Instruments, USA.

Change 11.043 DB2 PMSTA02 report count of SUSPENDS was usually zero,

ANALDB2R because not all suspend-count-variables were included.

Apr 27, 1993 Insert after line 138120 (reads 120 'QUANTITY';) this

statement: SUSPENDS=SUM(QTXASLOC,QTXASLAT,QTXASOTH);

and then change line 138180 from QTXASUS to SUSPENDS.

Thanks to Ed Lau, Lucky Stores, USA.

Change 11.042 DB2 PMACC02 report count of OPENs was actually the count

ANALDB2R of fetches (from the preceding line), because the 2nd

Apr 27, 1993 occurrence of AVG=QXFETCH /NUMPLANS; should have been

The second occurrence of AVG=QXFETCH /NUMPLANS should be

AVG=QXCLOSE /NUMPLANS; (in line 035330).
Change 11.041 ASUM70PR can corrupt PDB.RMFINTRV if you have a changed

ASUM70PR member IMACRMFI, because IMACRMFI only set the interval

TRND70PR for PDB.RMFINTRV. With this change, the interval of both

Apr 28, 1993 PDB.ASUM70PR and PDB.RMFINTRV are both set by IMACRMFI.

Insert after the comments at the beginning of the member

%INCLUDE SOURCLIB(IMACRMFI);

and after the ID = line, insert these two lines:

INTERVAL=MYTIME,

MYTIME = _DURSET,

which causes both datasets to have the same interval.

Variable NRPRCS was deleted from ASUM70PR (and also from

TRND70PR); it never should have been kept and its value

has caused confusion. ASUM70PR was revised to detect and

flag intervals during which the number of LPARs was

changed (DURATM and PCTCPUBY could have been incorrect)

wrong), and most importantly, these new variables were

added to ASUM70PR for enhanced reporting:

LPARCHG - 'Y' if something changed in any LPAR

LPnCHG - 'Y' if something changed in LPAR n.

LPnDUR - LPAR n's "Up time" or "Availability time to

execute CPU", the sum of DURATM across all

LCPUs in LPAR n. This is the total seconds

during which this LPAR could have been

CPU dispatched. If an LPAR was IPLed as

a 3-engine MVS machine, in one hour it

would have 3 hours of "Up Time".

LPCTnBY - Percent "Logically" Busy in LPAR n, equal

to 100*LPnUPDTM/LPnDUR, the Partition CPU

Dispatch divided by the LPAR's "Up Time".

If a 3-engine LPAR was dispatched for 56

minutes, it's LPCTnBY would be 31%. This

new variable describes how much the LPAR

was logically busy; the existing variable

PCTL1BY describes how much the LPAR was

physically busy. If this same 3-engine

LPAR was executing on a 5-engine machine,

it's PCTL1BY would be 18%, because that

LPAR was using 18% of the hardware while it

used 31% of the CPU time available to this

LPAR. See the example below.

LPCTnOV - Percent "Logically" Busy in "LPAR Overhead"

100*LPnMGTTM/LPnDUR, to describe how much

of the new LPCTnBY was in LPAR management

of this LPAR.

See also the MVS Technical Note III.7, Newsletter TWENTY-FOUR.
Thanks to Tom Parker, Hogan Systems, USA.
Change 11.040 Error "DATASET TAPEMNTS NOT SORTED" in WEEKBLD/MONTHBLD

WEEKBLD results because the BY list for TAPEMNTS should have been

MONTHBLD BY SYSTEM DEVICE SHIFT TMNTMTYP TMNTTIME; instead of

Apr 27, 1993 BY SYSTEM SHIFT DEVICE TMNTMTYP TMNTTIME;

Thanks to David Ehresman, University of Louisville, USA.
Change 11.039 Variables RDRTM,RDRDEVCL,RDRDEVTY were dropped from the

VMAC30 TYPE30_V and PDB.SMFINTRV datasets (because they are JOB

Apr 27, 1993 related, not interval related), but the deletion was not

noted in the Change log for MXG 10.10.

Thanks to Diane Eppestine, Southwestern Bell, USA.
Change 11.038 Variable QTXAIRLM was inadvertently omitted from the

ASUMDB2A SUM= list in the ASUMDB2A (DB2ACCT summarization) and the

TRNDDB2A TRNDDB2A (DB2ACCT trending) members. It is now added, and

Apr 15, 1993 the SUM= list was sorted & collimated for easier reading.

Thanks to Alan Keeble, British Steel, ENGLAND.
Change 11.037 The Total Read IOs was miscalculated on the Statistics

ANALDB2R Summary Report. It was actually the total Sync Reads but

Apr 13, 1993 should have been the sum of Sync and Prefetch Reads. Also

the calculation of Get Pages per Sync IO was omitted.

Find these five lines:

PUT @1 'TOTAL READ I/O OPERATIONS (RIO)'

@38 QB1TRIO 10.0

@49 QB2TRIO 10.0

@60 QB3TRIO 10.0

@71 QB4TRIO 10.0;

And replace them with these twenty-two lines:

QB1TOTL=SUM(QB1TRIO,QB1TPIO);

QB2TOTL=SUM(QB2TRIO,QB2TPIO);

QB3TOTL=SUM(QB3TRIO,QB3TPIO);

QB4TOTL=SUM(QB4TRIO,QB4TPIO);

IF QB1TRIO GT 0 THEN QB1PSYNC=QB1TGET/QB1TRIO;

ELSE QB1PSYNC=0;

IF QB2TRIO GT 0 THEN QB2PSYNC=QB2TGET/QB2TRIO;

ELSE QB2PSYNC=0;

IF QB3TRIO GT 0 THEN QB3PSYNC=QB3TGET/QB3TRIO;

ELSE QB3PSYNC=0;

IF QB4TRIO GT 0 THEN QB4PSYNC=QB4TGET/QB4TRIO;

ELSE QB4PSYNC=0;

PUT @1 'TOTAL READ I/O OPERATIONS (RIO)'

@38 QB1TOTL 10.0

@49 QB2TOTL 10.0

@60 QB3TOTL 10.0

@71 QB4TOTL 10.0;

PUT @1 'GETPAGE PER SYNC READ IO '

@38 QB1PSYNC 10.2

@49 QB2PSYNC 10.2

@60 QB3PSYNC 10.2

@71 QB4PSYNC 10.2;

Thanks to Victor Chacon, NYNEX, USA.

Change 11.036 Variables QWACARNL and QWACARNE count entries and exits,

ANALDB2R but DB2PM reports the number of suspensions, which is

Apr 13, 1993 one half of the data values. Thus, insert after 038520

QWACARNL=QWACARNL/2; and insert after 038620

QWACARNE=QWACARNE/2;

Thanks to Waldemar Schneider, SAS Institute Europe, GERMANY.

Change 11.035 This MXG utility to actually measure TSO response time at

CLTIMER your terminal was affected by SAS Version 6. A "STOP;"

Apr 13, 1993 statement must be added after the last "END;" statement.

Additionally, LS=80 was added to the OPTIONS statement,

also "'FOR TRANSACTION ' _I_" was added to the PUT for

cosmetic effect, and the sample CLIST in comments was

revised to use Version 6 DDnames SASLOG and SASPRINT.

If you have overlooked this utility, take a look now. It

should be made available to all TSO users so they can see

their own response time at their terminal to prove how

good real TSO response is (or isn't!).

Thanks to Freddie Arie, Lone Star Gas, TEXAS.

Change 11.034 "INVALID DATA FOR TAOSTYP" messages result because the

VMACTAO variable TAOSTYP should have been input at PIB1. instead

Apr 13, 1993 of HEX1.

Thanks to Jeff Burnett, Anixter Information Services, USA.

Change 11.033 Variable ACTDLYTM can have small negative (-.0010) values

VMAC30 because EXECTM is .01 resolution but ACTIVETM has .000001

Apr 13, 1993 resolution. After both calculations of ACTDLYTM, insert

IF . LT ACTDLYTM LT 0 THEN ACTDLYTM=0;

Thanks to Norbert Korsche, OMV-AG, AUSTRIA.
Change 11.032 Two new variables, TRANTIME and TRANCOST may exist at the

VMACMEMO end of the accounting segment, after EXITAREA, and are

Apr 13, 1993 now added to the KEEP= list and LABELed:

IF MEMOACTL GE 84 THEN INPUT

TRANTIME PIB4. /*transmission*duration*/