The Kodak quality package

THE KODAK

QUALITY

PACKAGE

J. L. Bossert and J. A. LaBarr Eastman Kodak Compuuy, Rochester, NY,

It was decided that the start-up team would consist of four members: the team leader, the programming and hardware coordinator, a quality consultant who would determine the fitness-for-use and a person who would develop a criteria, 11 criteria. strawman” based on the fitness-for-use Each person had some role to play which, when combined, would form a comprehensive data base to proceed from. The team leader was responsible for the overall coordination of the project. He would also be responsible for obtaining information on all types of quality systems and software. He would be the liaison to the Quality Advisory Committee and other organizations at Kodak. The programming and hardware coordinator would coordinate the activities of the programmers and provide information on hardware and software availability, accessibility, and compatibility. The quality consultant would be responsible for determining what programs were needed, the priority of the programs, what hardware was preferred, and user interest. The consultant who developed the strawman was responsible for the format of the package: how the package was developed -- modular? integrated? user-friendly? user menu-driven? command activated? -vicious? Then he would present what these types of things. he felt was most appropriate based on the fitness-for-use-criteria. So the team was picked; I was in the role of the quality consultant who would determine the fitness-for-use. Julie was the programming and hardware coordinator. What were the needs of the potential users? Many people at Eastman Kodak were enhancing their knowledge on statistical process control. What were the types of processes which would be attempted to be controlled, and what type of techniques would be utilized? Kodak Park manufactures many products; the processes vary as much as the products; there are continuous processes, batch processes, specialty and routine products. products, The initial task was to find the right combination of techniques for the initial package which would f.it the needs of the majority of people, Kodak Park had embarked on a training program where individuals within the various manufacturing divisions where trained to teach their divisions how to broaden their use of statistical process control It was decided that these individuals techniques. would provide a good sample of what was needed. Some things were obvious; X bar and Range control charts, histograms, and some attribute control charts. many people expressed an There was one surprise, interest in statistical inference. So the team did a Pareto analysis to come up with the three initial programs -- X bar and range control charts, and statistical inference. histogram, It was decided that after these were developed other parts would be included. Now the question remained: what system should this package be developed on? Much time and energy focused on the PC vs. mainframe question. PC’s were easier to program, but slow and not readily available. Mainframe packages were looked upon as

INTRODUCTION quality

We are here today control package

to talk about a statistical which we developed on our

IBM mainframe utilizing SAS; as our foundation. are proud of this package because:

We

1.

It shows the adaptability of its users.

2.

It was the first IBM mainframe quality ever developed at Kodak Park.

3.

It met the growing needs of Eastman Kodak as it searched for ways for improved quality in its products.

of SAS; to the needs package

The development of this package was anything There were times when KQP (as we but smooth. affectionately call our package) seemed like a This presentation is millstone around our necks. to provide a roadway for others on how to bring a quality computer package into existence. The talk is divided into four segments: Background, Development, Implementation, and Future Developments. BACKGROUND Eastman Kodak Company has a number of manufacturing units, the largest of which is Kodak Park located in Rochester, NY. At Kodak Park, there is a steering committee of upper management which tracks quality activities having multiple user In June of 1984, the Quality Technology Unit input. approached the Quality Advisory Committee for the support and funding of a statistical quality control This was given with the computer package. contingency that a preliminary package would be This ready for field testing by February 14, 1985. meant that the team formation, planning, programming, and debugging had to be accomplished Before any in approximately seven months. programming could take place, four items had to happen : 1.

A project

team had to be formed.

2.

The “fitness-for-use” determined.

criteria

3.

What type of computer system was to be utilized -- mainframe, PC, both.

4.

What type of programs product.

should

USA

had to be

be in the initial

1 SAS is a registered trademark of SAS Institute, USA. No endorsement is Inc., Cary, NC 27511-8000, implied by Eastman Kodak Company.

17

vicious, and having slow response during peak usage, but they were accessible throughout Kodak The final decision was to develop first on Park, the mainframe, and then downstream to consider PC The mainframe was chosen as the most applications. All the statistical efficient use of resources. programmers were well versed on the mainframe, a majority of the PC’s had communication links to the

-___-__-_-____--___

user

mainframe, and we could utilize SASias the foundation for all data handling and graphic capabilities. Members of the Quality Technology Unit then developed the specifications necessary for the programmers to work from. At the same time, our Our package was designed strawman was developed. to be modular in format; this way it would be easy Secondly, it would to add new parts to the package. be menu driven so that anyone could use the package regardless of hcu much or how little computer the “PC knowledge they had. This concept was called in the mainframe”.

*(pp

..->

Charts

for

Fixure

2.

Co”rrol

Chart*

_____-____--______

2

“ARIAOLE

cl.,.

Sample quality

package

control

chart

Once their selection is made, they receive menu specific to the chart type requested. In required to perform the Figure 3, the details individuals/moving range charts are requested:

i DEVELOPHENI’

OPTIUS

KQp)z‘,

____________

,(Qp

C,,.rt

of

,“di”,,,“o),

menu. the

_-____-_-__--___.___

CUmLA..\‘D --->

At Kodak Park, we use SASR

(1)

as a decision

the time of development of this support tool. package, SAS did not have any statistical process control (SPC) procedures available. We chose SAS/GRAPHR (2) to generate the graphics in this At

package because of its ability to *replay” plots. Replay capability gave us better control over directing plots to various graphics devices. Our initial intent was to provide some basic Our users have SPC tools -- no frills attached. Hany of the people varied computer backgrounds. attending our training program had never before It became obvious worked at a computer terminal. that a menu-driven front-end to our SAS SPC tool was An interface to direct plots and required. printouts to graphics and printing devices was necessary as well. To give you an idea of our menu structure, let’s look at a couple of package menus. When users first enter the package, they receive the main menu:

__--___-____-__-___ K&lb

uP?:o.s

.I.>

pu,],CY

p,&,g*

-____-____-____----_-

2

From here, they select an application; for They then receive a menu example control charts. that allows them to select which type of control chart they want; for example a chart of individuals:

2 SAS/GRAPH is a registered trademark of SAS Cary, NC 27511-8000. USA. No Institute, Inc. endorsement is implied by Eastman Kodak Company.

.b”ge*

upper l.i.*t A,. Lower

3.

Sample quality

package

chart

of

L,.,f

individuals

-.t ..> ..>

menu.

Our first release of the package included histograms, X bar and range/sigma control charts, and statistical inference. We decided to place these tools iu the hands of the users -- we wanted feedback on their likes and dislikes of the package and what other tools were needed. KQP was a “prototype” system. Before releasing the package to the general public, we selected 15-20 people who expressed an interest in testing the package. These were people from all sections of Kodak Park. They had varying levels of computer experience, but none were computer novices. They tested our product for approximately six weeks. Any bugs detected were fixed. We received feedback on likes and dislikes of the package. For example, concern was expressed over: 1.

The automatic plotting of the normal distribution on the histogram, and

2.

The selected pattern style for the control limits on control charts with unequal subgroup sizes.

One need became apparent -- a data entry/modification feature. Some of our testers were SAS users, some were not. Those unfamiliar with SAS needed help getting their data into a format KQP understood. Another need also was identified -- a feature allowing users to store and replay plots and associated statistical output at the screen and then replay to hardcopy devices. Another need also was identified -- a feature allowing usera to store and replay plots. An ideal way for users to run KQP is to display the plots and associated statistical output at the screen and then replay to hardcopy devices. Our first attempt at data entry was unsophisticated, but answered the immediate need.

19

The Kodak QualityPackage From the UTILITYsection,users selectdata entry and specifya datasetname:

_~~____________~~~_~~~~~~~~ xQp.9PgtflltlC*

OPTlOS --->

multivariatecontrolcharting-- specifications were package , drawn up. We went from a three-application in Spring 1985 to eight applicationsby the Fall. Modifications to the current modules were also made:

___.__________________.

1

Controlover cell size in the histogrammodule was added.

DATA E?iTRS;xASIPt’LATIOS I Create II IICW SAS dstsrtt 2 Edlt a SAS dataret 3 InteraCtIVe SAS

7 S

:

The limit calculationsin the X bar and rangefsigmachartswere changedfrom the look-up to algorithms. This providedus with more flexibility.

Print current listing fl1a Replay plots InI(PP 7 PLOTS (fIlename filstypa fIlemode)

i SiS datasct

sample

4.

The overlayof the normal distributionon histogramsbecame optional.

name -->

DEMO (Ilbraryl

. KGPUATA (for (dataact

options 1.2.4.

and 5)

quality package utilitiesmenu.

They then receivea screenwhere they define tbeir datasetvariables:

The need for plot and listingoutput consistencyamong modulesbecame apparent. Specifications for linepatterns.colors,and symbol types were written. We had no written documentationfor KQP -insteadwe includedan on-linetutorial. The tutorialis availablefrom either the main menu: WOd&

___“_______----_---

-.-a

OPTlOS CREATE

___-________-____-

couxAso

Create

,

$A$

Length (Char var) 1 only I l **

TYPC IS/Cl *

h’ane ..*..*a. 1 f3SC 2 DEXSITY 3 BATCH

for

each Varltblr

Yw

---------------------

C I X

VISCOSITY

s B C

USERID TllE TEWSXAL PF KEYS

Label ..~**I.~.*~..*.....~*.~~~~...~*..

CUS Comaand Tutorial Exit

-

123456 cl*:01 3218 24

!

or Excc%

Figure 7. Sample quality packagemain menu.

HAWFACTURIXG SATCH BCXUER

G

or from an applicationscreen. For example,if users are runninga test of means for sample vei-sus populationand do not understandhow to respond, they simplyenter"HELP"on the "Commande" line:

: 6 7 8 0 j 10

: :

5. Sample quality package variable definitionscreen.

)‘Qp320

Edit

SAS

data

ret:

OEIG.K”PDATA

-.I>

: Sercc” ---------

1

SAS

DESSITY:

Figure 6.

_

datattt

KQP

st*ti*tic*1

name --->

--->

DEXO

*---..------------

IWPUATA ldstssetl

“ISC

j

Variable

Standard Value (~0) Alpha Level Altrrnate Hyjmtkcrls

I

*nf*rc”cc

KELP

’

I : I

v1sc:

--r>

11Ibrary)

______________.____*_________________________________~_~~~____~

BATCH:

.._.._--------__--

~o?I!+As3

Once their variablesare defined,they enter data, one observationat a time:

Comand

P.&age

want

*

pipure

pua1,ty

T

__________s______--.

--->

Enter rkc name. type. ltnnrk, and label . KUPDATA on the new dateset DE!lO

j

d,c,se(

Last -..> I..> -I.>

#OS A - Sot equal B - Lets than C - GrtW.er than

Ll8tsng OUtput: DISPLAY DO you want te change

this selection-I> Ko

Figure 8. Sample qualitypackage statistical inferencemenu.

Sample quality package data entry Pcrech.

SAS's full-screeneditor (SAS/FSPK(3)) provides this data entry capability. IMPLEMENTATION After testingwas completedand the data entry featuresadded, KQP was releasedto the general public. It remaineda prototypesystem. We wanted our users to tell us what more was needed. At Kodak, areas workingwith chemicalscannot always test in logicalsubgroups. The need for charts of individualswas critical-- hence our control chart option startedto expand. Specificationsfor P/NP charts,C charts,U charts,and Pareto analysiswere also written. One area expressedan interestin

3 SAS/FSP is a registeredtrademarkof SAS Institute,Inc., Cary, NC 27.511-8000, USA. No endorsementis impliedby Eastman Kodak Company.

Queued:

and they are placed into the tutorialfor statisticalinference:

_______-____________---____ xpp &jIj~lC,

---a

OPTIOK DATA

__---__--__----_--.__

2

ESTRY/IUSIPUL*TIOS 1

crea,*

* 3

EdlC a S&S datasrt I”tCract,Y* IA.5

. new

DATA/VARIABLE I Rror,e S

s*5

OPTlOSS . SAS d.t.,Ct

Browse

PLOT/LJSTISCs

SAS *,t.ret

VarJablt

IJat

ol

.

SAS d.t.wt

OPIIOSS

data*ct

n*.e

--a

DEW3

KQPDATA

(for

09tJom

1.2.4.

and

I,

lJlbr.ryJ’ld.t...t)

Figure

1O.B

Sample

quality

package

utilities

menu.

You enter the appropriate option and dataset name. From there you are placed into SAS’s full-screen editor and you start adding observations to the dataset: EdJt C0...“d

SAS d.t.

set:

) scrren ;_________

DL~O.%QPMtA

--->

: Se”

1

12,

_________________________________________----------------------

SIll9lC

!

man

-

Z.T.Chl-Sq”ars.F

-

I

Figure

the

calculated

the calculated co.p.red to rJSnlfJCanCe.

smyle

the

117

DESSITY:

130

UATCII :

J-102_

avsra~c

tc,, rtatJ,t,c. crltlcol value

Sample qua1 itv package inference tutorial.

9.

wsc:

statist

Th,, value to detrraln*

1,

Sample quality CIIt ry screen.

1O.C

Figure

._ ________________ ,(dak

To enter the package, enter “KQP”. You receive the KQP main menu:

OPTIOS

.-.>

lo**k

Q”al‘ly

p,c*,~c

Figure YOU

,,“,lJfy

P.Ck,ge

a

---------------------

1

Sample

10.D

receive

the following

quality

package

main menu.

the histogram application specifications:

menu and make

I.

The variable

you want a histogram

2.

The subgroup

variable,

3.

Historical

4

specification

limits,

Titles. I(pp,ao_--_______--_________kpp “J‘tolr,.

of.

and

____________-___---.

co.muso --->

PP KEYS

10.A

---a

O,‘,,tcS

4.

_________-_____-----.

USER10 TIME TERM*S*L

l.i?ure

data

After adding this week’s data, you would like From the main menu, you enter the histogram. HISTOGRAMsection:

ical

Our tutorial was not enough. First-time users Many of requested demonstrations of the package. our users had used our computer facilities previously only for electronic mail -- or not at all! At our on-line demonstrations, we walked through how each of the applications function in KQP, and how to enter data. The demonstrations and tutorials provided users with enough information to get started. Computer and SPC consulting for the package is also available. A user manual is provided with the Fall release of our product. This manual contains a walkthrough of the package. Making the entries shown and using detasets we provide on our system disk, users can get a feel for how KQP functions. Let’s go through an application in KQP from start to finish as a real user would.

___-____ ._________-

package

Sample quality

package

-

122456 09:01

- 3218 -24

main menu.

1 j

Rrr9onsa S”bSro”9 Start 2nd

: ;

You have an existing this week’s data to, section:

database that you want to add so you enter the UTILITIES

c99.r AIn Lover Normal

vnrlablc varloble subSmu .ubSroup

.-a DEXSlTY --, MICB awb.r I.> number -.,

spcc

1.9 12s 110

S9.C Cur”.

O”.rJ,y

C.J, c*JJ C.JJ .I,

/

Title Subtltlc

j , i

ora9hlcr o”t9”t: DISPLAY LIstlnS OYt9”t: DISPLAY 90 YOU “.“t to ch.ng. them

Figure

-I> -.a

--a --a -->

10.E

‘a,(

“.xJ.“. SJre MI”‘.“.

Queued: rot.,:

0

I.> -.a -.a

“0

KQP NALKMROUGI ,SI,lULATED DATA,

Sample

IBXJPW s.,.ctJon.

quality

I->

package

RO

histogram

menu.

21

J. L. Bossart and J. A. Labau

; __________________xppCrr"tror Ck.rt* -------_r---_------! I OPTIOX ...a 1

You will note et the bottom of the screen that the listingand graphicsoptions are set for DISPLAY You submit this requestand view the histogram:

, j Charts

for

Xbar snd Ran2elSiS~~ Chart IndlvIdualr and Kovtng Range Chart CCXSL‘RclI*rt

1 2 3

:

1

VARIABL2 data

! Charts

for

1 :

KP Chart C Chart P Chart C Chart

4 3 6 i

/

ATTRIBUTE data

Figure 10.1 Sample qualitypackage control chirt menu. You receivethe X bar and range/sigmachart applicationmenu, and make the following specifications: 1. The variableyou want a control chart of, 2. The subgroupvariable, 3. Historicalcontrol limits,and 4. Titles.

Figure 10.F Sample quality packagehistogram.

I(pp2,o

I(QP x&r Chart _____-r___*__________ :

___________________

CORRASD .-->

and associatedstatisticallistingoutput.

SAS d&taxer KPP WALKTKROV6H (SlWULATEO DATA) HirtoSrm

nsne 1-1

Rcspon~ variable

I

-.> ..s SubSroup variable St.rt wbgroup number ..> subgroup number ..) End

I STATISTICS OF PLOT - OE~SITY

KQ~DATA ‘fd.t.srt)

Dii*O fllbrrry)

DEKSITV 6ATCR

! s - 120 xtx = 108 nAx* ml nus - 12'1.6 ST0 L)EV - 6.716594 WV - 5.257666 KLM l 2 Srr, k’ . 147.6568 K2AS - 3 ST0 OEV . 107.6412 AlI = 125 LSL * 110 SBELW LSL - 0.4026676 USL - 140 * ABCZVE CSL * 3.471666 TOTAL L OLWIDE SPtC - 3.815384 P.C.I. - 0.6051953

Plot

type

Ca’cufatcd

==) KASGE Lfaltr:

Slgmb ifnits

-->

litit Subtltle

*I> -->

Grsphlc. Llrtlng

Output: OUtpUt:

tlkngc

ot‘ STD)

nems LimIt --> AI, --> Lover LInlt -.a

!Enttrcd:

!

3 : :

upper

I.70 I27 I24

STDlR4nSer --> Al. I-> Lwn?r L‘mLt ..>

upperLIDIt

: ’ j :

KQP IIALKTKROUGH (SIRULATED UATA) DISPLAY - lBR2279 DISPLAY

Figure 10.3 Sample qualitypackage Xbar/range chart menu.

I

Figure 10.G Sample histogramlistingoutput.

You submit this requestand view the controlchart: You now want an X bar/rangecontrolchart of the data. Again from the main menu, you indicatethe CONTSOLCHARTS section: Rod&

__________L-----“-_

Q”@,,ty

p*&,gc

__________-_--_r”Lr””

:

OPTIOS .--a 2

, : : ; :

1 2 2 4

USERIO TM2 TERRIKAL PF KEYS

- 123456 - W:lS - 2276 - 24

i : j

c 1 x

CUS Command6 6t Tutorial Exit

Ixcct

Figure 10.11 Sample quality package main menu.

You next select the X bar and range/sigmachart from the CQNTSOLCHARTS selectionmenu:

Figure 10.K Sample qualitypackageXbarfrange chart.

and associatedstatisticallistingoutput.

KQP w*LKl?llO”GR

I

lSIwL*TED SCSGROI’P

DATr\, POR T”E “Mlr(E.LE

IxrORUTIOS

DESSITV

9 j

SUBGROCP

S”BGROUP

SlmxOCP

X-SAR

I.D.

hZ’“BEK

SAXPLE SIZE

I

A-101

! I !

A-102

! : :

W-102 w-103 1-101

X-SIR

x-SAR

X

LDVEK

CPPEK

LIXIT

Ll”,T

LIMIT

EXCEEDED

WUBLE BAR

124

,30.0

130

121

i

12‘ I24

12L.Z

130 I30

127 127 12,

,

127 127

I

CPPER LDmK

127

1 2

5 S

3 4

S 3

22 23

5 S

12. 124

127.4 133.0

130 130

24

5

12.

111.0

130

A-103 B-101

120.4 129.4

124

KQP SCSGROCP i

X-SAK

SUSGROUP

SCSGKOL’P

RASGE

I.D.

BlnSER

SMPLE SIZE

; ,

’

XALKTKKOUGR

~S1UCLATF.G D*T*, TOR I”2 Y*R,*SLE

,twGK..T*Ds

SCSGRGCP

i :

130

RASGE

DLKSlN MSCE

R*sGE

KCISGL

LOW9

UPPER

LIMIT

SAM

LIXIT

Llllil

EXCEEDED

A-101 A-102 A-103 S-101

1

S

0

3

11.214i

5.33333

2 3 .

S 5 S

0 0 0

1 1 3

Il.2747 11.2741 ll.2747

1.33333 5.33333 5.33333

:

n-102 K-103 I-101

22 23 24

S 5 S

0 0 0

0 S G

11.2747 11.214: 11.2741

0.33333 5.33333 5.33333

v

Figure

10.1.

Sample

Xbar/range

chart

listing

outpllt.

Now that UTILITIES

you have section

you would like done these analyses. of KQP again from the main menu and

a hard select

copy of the control chart. the REPLAY PLOTS option:

You

-___--__ - KQP Ct,,it,e, ____-___-________________ j __________-____--OrTlOS .--> I

i

1 PLo7,LISTIno 6 *ro*so

I MS

d.t.wt

na.c

10.M

k’igure

A screen

appears

requesting

OPTlOSS currmt

where

you

------------

Li,,‘“S

D”t9”l

I ’

Valid

09tion pr1ntcr: pr,nt.r

option

KQPDATA (dataset)

quality

like KQP

the oUl9”t

a

P or

S.

I0.N

utilities

-----

VERIPV

output

Valid

Print

i If

51

;

menu.

0

OKAPHICS

0ptlon

Grr9hIc

LGCATIOS

I

--.>

09tlons:

D G 2 I

-

DIsplmy Could fin 2.t. (I” IM 32Sl

9

-

pr,ntcr

Option Dwic*

- “. ..I)

09t1c.n Xodsnme

-

BS81 S3S) or co.mtlblr

’

plots spcclfy

device:

;

I I(PCC *wmEnn oodsna.c

Figure

09tlons

GT.9hlC

--->

and

sent:

plots

--->

spsclfy

oPtlone1.2.4.

(for

package

09tlon8:

Dis9lay Print Dls91av

-

f1le

OL*O (library)

Sample

L‘stinS D P I -

If

..a

would

KQPPIOF

lIstinS

Sample menu.

, I

quality

If

package

I. spc‘Iy nodenmr -1-> KPOO1001

output

options

,

reenter

the

23

J. L. Bossart and J. A. Labau After selecting a device, you select which plots you want a hard copy of:

Grsphlcr : Coamand

Replay

The package prompts you for a filename for storing your plots and returns you to the host operating system:

Procedure Enter

--->

! h’otc:Forhelp,

’ Plots

j R: _ PROC GPLOT OUTPUT _ PLOT OF Y’X

;

Figure 10.0

a new name

15AUGBS

09:12

lSAUGB5

09:lT

Iron

this

Figure 10.S

PROC GPLOT OCTPLT PLOT OF S’SCBGROCP

for

the

KQP plot

file:

ASQC

press PF 1 or PF 13 ___________________________________-____________________________

scrrlon

have

been

saved

In

ASQC PLOTS A

j I

Exiting the KQP Package.

FUTURE DEVELOPMENTS

Sample quality package graphics replay menu.

Based on input from our users, many specifications are being developed for future The plots are directed to the selected graphics inclusion into KQP. A sampling of these location. You also want a hard copy of the specifications are: median control charts, CUSUM statistical listing output. From the UTILITIES menu control charts, normal probability tests, power you enter the appropriate option: curves for sampling plans, a sampling module, and bivariate control ellipses. The user community ____________________-______ KPp L'\,l,tlet -_______________________ submits requests in writing, and we act on them i OPTIOS ---> 7 based on input from the Quality Technology Unit at Kodak Park. DATA ESTRY/MASIPULATIOS a new SAS dataact

1

Create

2 3

Edlt a SAS dataget Interactive SAS

: DATAIVARIABLE ; 4 Browse

OPTlO!iS a SAS dotaset

:

Vorlablr

5

Browre

list

PLOT/:ISTISO OPTIOSS 6 Browse current j

i

1

currenr plots

Prlllt Replay

8

] SAS dhtaset

nam.? -->

I

of

lIstIn

a SAS dataset

fllc

11tt1ng file in KQP (filename DWD

PLOTS tiletype

T flIernode)

KQPDATA

(for

optjonr

1.2.4.

and

5)

(1Ibrtlry)‘ldatasct)

Figure 10.P

Sample quality package utilities menu.

and again select the output destination.

, KQPPROF

------------

KQP Output

Options

--------

VERIFY

PRIRT

Option

--->

LOCATIOS

,

I

: I

output option ---,

Lirtlng Valid

Listing

D P B -

j

If Optlon printer: Printer

Display Print Dlrplay

-

Figure 1O.Q

i

Graphic

I

Options:

h Print

P or’ B. ---,

P

specify

REMOTE7 KPCC RENOTE~W nodcname

Output

Valid

Graphic

Options:

!

0 t -

Dlrplay Gould

2 - zeta (Ill RS6) 1 - IBM 3267 or compatible

I

A F P -

AGlLE connected to p!ottsr Save In plot file only printer plots

j 1

If

Optlon Dtvlcc

j

xf

option Sodename

1

- 0. --->

(In

856)

specify

I. rpeclfy .I.>

devlcc:

nodename

Sample quality package output options menu.

The statistical listing output is directed to the selected location. You have concluded the analysis you want to perform, so you exit KQP:

!

K,,,j&

_________.__..r___-

1 OPTIOS

I

---,

I

; 1

1 Histogram 2 Control Charts 3 ftarlsticml lnfercnce , UtlIlticr

:

X

1

P.&age

___________________--

USERID TIKE TERMINAL PF KEYS

-

123456 OS:20 3276 24

ErIt

Figure 10.R

RTP-C

Quality

X

Sample quality package main menu.

)