Sider Scan uses a proprietary algorithm developed in-house to find duplicate code with maximum speed. It then uses its comparison engine to quickly determine the quality of duplicates. The following is the general process of analysis.
1. Group the duplicates
Sider finds duplicate blocks of code in your project and groups them.
2. Pattern analysis per group
Sider Scan performs two methods of analysis.
(a) Pattern-based analysis
For each pair of duplicates, a diff library is created and pattern analyses are initiated to determine if there are any problems. This is referred to as the 'Pattern' method of analysis.
(b) Action-based analysis
When Sider Scan analyzes code over an extended period of time, any modifications/additions/deletions made in the code will be tracked, and problematic changes will be part of the analysis results. If a pair of duplicate code is detected, and when a modification occurs in one but not the other, these will be flagged as potential problems as a 'possible miss'. This is referred to as the 'Time-series' method of analysis, and is specifically based on actions taken by a user.
Please note that this time-series analysis is only possible when the scan is consistently run at regular intervals.
3. Notifications on the potentially problematic code discovered
An email notification with the analysis results on problems or pattern irregularities is created, where details and suggestions are visually presented for you to review.