New Method is found to make the Programmers’ Life Easy
This is the blockbuster news for the people related to Software Engineering that a programmer writes a program without bugs (errors). It is just like snowfall in Karachi. Program and bugs are linked together. There still remains some bugs even if thousands of people and brains struggle together with much care to make a program. For the solution of this problem, many Algorithms are used to find these bugs but benefits of these algorithms are limited. Massachusetts institute of technology has introduced one more use of Machine Learning. Machine Learning can be used to correct any code full of bugs. Researchers at MIT have claimed to make such a system which will correct bugs of any source code more than 10 times than any system. The researchers named this system as “Prophet” in their research paper.
This system makes itself better by analyzing the correct bugs of open source programs. It is said that “Prophet” is a method for open source programs to patch 777 bugs of 8 open source applications on well known “Git Hub”. The system with this information not only searched the bugs but also patched these bugs. Fanlong the author of this research and graduate of Electrical Engineering and Computer Science, has developed a program which can correct the bugs of program code. Fanlong’s software has only one fault that it takes very long time to do its job.
Experts checked this system on 69 bugs of real life which occur in 8 open source programs. The bug repair system available before “Prophet” can patch only 1 or 2 of the bugs but this new system can find 15 to 18 bugs and successfully can patch these bugs.
According to Martin Rinard, professor of Electrical Engineering and Computer Science, and co-author of this research piece, this new system is defiantly very useful. According to him, it is found from this research that a correct and bug-free source code has some qualities which we can learn from a program and then can apply to some other program.