Semi-Interpretive becomes more thorough and efficient; then it takes over.
For the fourth following semester, Spr14, the provost wanted schools to have the ability to process two semesters concurrently. The provost realized that during the fall, some students were applying for spring and some for the next fall. The provost also wanted statistics. Engineering had a much higher ratio of rejections than Fine Arts had.
The Error
The assistant to the provost detected the cause of the higher ratio. In the table for Engineering, the row with score = 0 catches all of the applications with scores less than 50. In the chart for Fine Arts, any score less than 50 falls thru to the exit; no action is taken. This error existed for three semesters and several reviews. The Provost was suspicious of the School of Engineering, but his own assistant found the error. The error explained several letters from applicants who complained about a lack of acceptance decisions.
As for the concurrent requirement, the program in Engineering handled multiple semesters concurrently from the beginning, as was demonstrated in the second semester. The program did require the input of the semester key. The program in Fine Arts would have required significant doubling of its logic, and then the developers realized that for each semester, the program required much more maintenance and testing to install two semesters. Consecutive semesters usually had different flowcharts. On the menu system, the system admins in Fine Arts had been keeping a version of the program for each semester; hopefully, the clerk chose the correct menu choice.
Fine Arts again took 11 days including one to find the big bug for a total of 41. Engineering spent 2 hours adding table entries and a day of testing for a total of 12 days. Engineering did not need to install.
By the next semester, the Provost tweaked the specs a bit. The programmers in Fine Arts estimated another 6 to 11 days to modify, test, and install. Engineering estimated one day for table entries and testing. The programmers in Fine Arts recommended adopting the program from Engineering.
FIne Arts about 41 days, Engineering about 12 days.
This ratio for maintenance days is typical when comparing typical methods to semi-interpretive methods. Finding real life comparison proves difficult. Most CIOs seem to shy away from scrutiny, especially when inspection might reveal great in-efficiencies. During the 1980's, Price Waterhouse reviewed the Federated Group and compared it to Circuit City and to Silo. The accountants at Price Waterhouse said that the Group had more functionality and better controls than the other two. The Group had five programmers and Circuit City supposedly had over two hundred. Such a ratio is so extreme that it begs for confirmation. See Software Research Northwest
The Error
The assistant to the provost detected the cause of the higher ratio. In the table for Engineering, the row with score = 0 catches all of the applications with scores less than 50. In the chart for Fine Arts, any score less than 50 falls thru to the exit; no action is taken. This error existed for three semesters and several reviews. The Provost was suspicious of the School of Engineering, but his own assistant found the error. The error explained several letters from applicants who complained about a lack of acceptance decisions.
As for the concurrent requirement, the program in Engineering handled multiple semesters concurrently from the beginning, as was demonstrated in the second semester. The program did require the input of the semester key. The program in Fine Arts would have required significant doubling of its logic, and then the developers realized that for each semester, the program required much more maintenance and testing to install two semesters. Consecutive semesters usually had different flowcharts. On the menu system, the system admins in Fine Arts had been keeping a version of the program for each semester; hopefully, the clerk chose the correct menu choice.
Fine Arts again took 11 days including one to find the big bug for a total of 41. Engineering spent 2 hours adding table entries and a day of testing for a total of 12 days. Engineering did not need to install.
By the next semester, the Provost tweaked the specs a bit. The programmers in Fine Arts estimated another 6 to 11 days to modify, test, and install. Engineering estimated one day for table entries and testing. The programmers in Fine Arts recommended adopting the program from Engineering.
FIne Arts about 41 days, Engineering about 12 days.
This ratio for maintenance days is typical when comparing typical methods to semi-interpretive methods. Finding real life comparison proves difficult. Most CIOs seem to shy away from scrutiny, especially when inspection might reveal great in-efficiencies. During the 1980's, Price Waterhouse reviewed the Federated Group and compared it to Circuit City and to Silo. The accountants at Price Waterhouse said that the Group had more functionality and better controls than the other two. The Group had five programmers and Circuit City supposedly had over two hundred. Such a ratio is so extreme that it begs for confirmation. See Software Research Northwest