EUGENE P. URETZ, JOHN A. MCCLARY AND STEVE MORGAN 
1253 
S22 URETZ, E.F. 
141037 = SGPT « 
ACODE? A123J 
4 
CODE FILED 
14 
VDftTE? 
4 
XTINE? X1B15J 
CODE FILED 
CONDITION 
= SH 
FILED U 
HIGH O.D. 
.275 
FILED i* 
LOU O.D. 
.112 
FILED 1* 
OIL. FACTOR 
1 
FILED i« 
KEYIH » TO PROCESS t_ 
Figure 13. — Entering test continuation request. 
specification shown in Table III for test T41037. 
It only remains for the user to key in # as in- 
structed. (Figure 13). This is a special symbol 
which indicates that the test should be contin- 
ued. In the case of T41037 this means calling on 
a processor to perform a simple numeric calcu- 
lation with the numbers. Since the calculation is 
a simple one requiring only milliseconds, the 
T41037 processor has been stored at the highest 
priority user processor level. This processor is 
loaded, the processor looks up the keyed in data 
values in the DAYFILE where they are stored, 
computes the value of SGPT, stores the com- 
puted value in the DAYFILE and generates a 
display which gives the SGPT result (Figure 
14). 
The laboratory technician may then key in an 
! which causes a new test cycle to start and the 
display shown in Figure 9 is again generated to 
guide the next T41037 test. Alternately, he may 
key in a new test number, or the symbols T9999 
which is another special symbol which signs 
him off the terminal. He may also key in a — 2, 
which is another special symbol. This value is 
stored in the DAYFILE and instructs the re- 
mote system to delete the series of data just en- 
tered. 
Response Time Considerations and 
Priority Structure 
The priorities to which the various system 
and user processors have been assigned were 
carefully selected to give response times com- 
mensurate with the complexity of the job being 
performed. Consider the example of the pre- 
vious section. The dialogue which occurred be- 
tween the lab technician and the user was under 
control of an executive program. This program 
is expressly designed to handle all dialogues 
with all terminals essentially simultaneously. 
Thus, no single terminal is delayed more than a 
few milliseconds because of dialogues occurring 
with other terminals. Furthermore, the execu- 
tive routine as well as terminal reading and 
writing takes place at a higher priority than all 
of the user processors, so response to a key in 
will not be delayed by slow user processors. 
Also careful assignment of processors to the 
four priority levels available insures that a fast 
computation will not be unduly delayed by a 
slow processor which will be assigned a lower 
priority. Furthermore, sampling takes place at 
the highest priority available and will not be de- 
layed by any other foreground or background 
processing. 
Analog Sampling 
The example shown above did not require the 
acquisition of analog data. However, some tests 
do require sampling and analysis of large 
amounts of analog data. A SAMPLER program 
has been developed to accommodate this func- 
tion. Sampled data is placed in a double buffer 
S22 URETZ, E.F. 
T4ie37 : SGPT 4 
ftCODE? A123J 
4 
CODE FILED 
14 
YDATE? 
4 
XTIHE? X1815J 
4 
CODE FILED 
14 
CONDITION 
= SH 
4 FILED 14 
HIGH O.D. 
.275 
4 FILED 14 
LOU O.D. 
.112 
4 FILED 14 
OIL. FACTOR 
1 
4 FILED 1« 
RESULT= +.27167E 
-81 KARHEH 
UNITS « 
KEYIH • TO CONTINUE TEST 1 _ 
Figure 14. — Display of test results. 
