1.if !\n(xx .so tmac.p 2.nr H1 3 3.if n .ND 4.NH 5Conclusions 6.PP 7It is appropriate to consider, 8given the amount of time invested in rewriting the interpreter, 9whether the time was well spent, or whether a code-generator 10could have been written with an equivalent amount of effort. 11The Berkeley Pascal system is being modified to interface 12to the code generator of the portable C compiler with 13not much more work than was involved in rewritting 14.I px . 15However this compiler will probably not supercede the interpreter 16in an instructional environment as the 17necessary loading and assembly processes will slow the 18compilation process to a noticeable degree. 19This effect will be further exaggerated because 20student users spend more time in compilation than in execution. 21Measurements over the course of a quarter at Berkeley with a mixture 22of students from beginning programming to upper division compiler 23construction show that the amount of time in compilation exceeds the amount 24of time spent in the interpreter, the ratio being approximately 60/40. 25.PP 26A more promising approach might have been a throw-away code generator 27such as was done for the 28.SM 29WATFIV 30.NL 31system. 32However the addition of high-quality post-mortem and interactive 33debugging facilities become much more difficult to provide than 34in the interpreter environment. 35