You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
TraceBoosting, as sketched conceptually in above figure, is an algorithm designed to enhance retroactive feature tracing with proactively collected feature traces.
15
-
Particularly, this implementation compares multiple product variants which are represented as artifact trees.
16
-
It builds sets of co-occurring artifacts by matching the trees and computes sets of possible and impossible features for these co-occurring artifact sets based on the respective configuration in which they occur.
15
+
Particularly, this implementation compares multiple product variants which are _parsed_ into artifact trees (V1-V3).
16
+
It builds sets of co-occurring artifacts (CA1-CA5) by matching the tree nodess and edges and computes sets of possible and impossible features for these co-occurring artifact sets
17
+
based on the respective configuration in which the tree elements occur.
17
18
This heuristic retroactive comparison-based tracing method is inspired by the algorithm used in the tool [ECCO](https://jku-isse.github.io/ecco/).
18
19
19
20
To increase the accuracy and efficiency of the heuristic algorithm,
20
21
we integrate proactive knowledge (represented as rectangles attached to artifact nodes in above figure).
21
-
Each time a proactive feature trace is available in a co-occuring artifact set and not contradicting another proactive trace,
22
+
Each time a proactive feature trace is available in a co-occuring artifact set and does not contradict another proactive trace,
22
23
the expression of the proactive trace is mapped onto all artifact nodes in the set.
23
-
In this way, we can propagate the reliable knowledge, provided typically by humans, to the remaining artifact nodes in a co-occuring artifact set and eventually,
24
-
spread the information across variants.
24
+
In this way, we can propagate the highly reliable knowledge of the proactive trace,
25
+
provided typically by humans, to the remaining artifact nodes in a co-occuring artifact set and, thus, essentially
26
+
spread the information across variants overriding the heuristcally determined annotation.
25
27
26
28
## Dependencies
27
29
-[Java using JDK17 or newer](https://www.oracle.com/java/technologies/downloads/)
0 commit comments