The report describes improved algorithms within a computer program for identifying spelling and word order errors in student responses. A "markup analysis" compares a student's response string to an author-specified model string and generates a graphical error markup that indicates spelling, capitalization, and accent errors, extra or missing words, and out-of-order words. The algorithm determines whether the response was acceptable or not, and computes a string of graphical error marks to be displayed below the student response. Synonyms and ignorable words can be specified and spelling errors, extra words, and word order errors can be accepted at the author's discretion. Spelling analysis is done using a dynamic programming algorithm that produces a least-cost edit trace; word order analysis is implemented using recursive branch and bound search. Improvements on earlier versions of the algorithm give more intuitive markup values. The algorithm is implemented as a HyperTalk XFCN. HyperTalk scripts can provide numerous input parameters that control the details of the matching process, and the algorithm returns a variety of fit measurements that characterize the match. Non-roman linear writing scripts are supported. The report contains detailed information on use of the function and serves as a user manual. Contains two references. (Author/MSE)