| In program language examination,Auto-marking of program is the key and difficult to realize the no-paper test.Now, Auto-marking system mostly have the character of ignoring the inward structure,and the result of scoring is unaccurate,no perfect or lack of flexibility.Althogh many experts and scholars are bending themselves to find a interinform and attempt to use such a form to denote the standard answer and to match the examinee's program .But,This method is to manipulate the program as the whole to compare and non-flexibility.By researching the correlation materials of internal and abroad,this text put forward the method of breaking down the score of progrm as four parts,that is, grammar score,testing score,knowledge key score and result score. according to the above-mentioned score's structure,Auto-marking system is divided three models:First, As some examinee's program can compile and the result of running is accurate, the system may give a mark by function testing to test the algorithm and programm's function,reference the result of testing.Second, As some examinee's program can compile but the result of running is erro,the system call the knowledge key scoring model to analyse the program,that is ,the program is divided several independence knowledge keys and described by regular express, after that searching for them in the examinee's program.Last,As some programs can't compile,the system call the grammar analysis model to give grammar score by lexical analysis and grammar analysis;then call the knowledge key scoring model to search the knowledge keys,and give the knowledge keys score by the result of searching.This system has already entered the test stage.The test result indicated that this system had already can run steadly and unify grading standard,the result of auto-marking is similar to the manual judging,and the process of grading pays attention to the internal structure of program,improvs the fair,the rational and the accurate . |
PDF Full Text Request