By alev
via javadots.blogspot.com
Published: Nov 30 2008 / 15:21
This is the last piece in a three-piece mini-series describing the flaws inherent in static dependency analysis (e.g.: Structure101). The author shows, by a concrete example, that dependency analysis can easily produce irrational results, thus making the whole approach quite useless.
-
For vs Each in Ruby
-
MapReduce Patterns, Algorithms, and Use Cases
-
Wyse CEO to Discuss Leading Trends in Mobility and Cloud Computing at the Stifel Nicolaus Technology & Telecom Conference 2012
-
Types in Java
-
A business component architecture with Spring 3.0/3.1 – Part 3: Properties
-
Corona SDK: Create a Rapid Roll-like Game – Setup
Comments
andrewm replied ago:
the article and reasoning is poor.
of course, structural analysis is only one part of quality in s/w. the small mutually-cyclic example presented can be solved using an interface (one or both ways) which fixes the structure nicely. can you obfuscate such an example to use reflection and thereby fool the static analysis programs? sure. have you fixed any of the underlying structural and coupling issues? no.
i'm a very happy customer of s101 and i've used it on large and small codebases to very good effect.
pulesen replied ago:
agree with you, I do
sutts replied ago:
I do feel duty bound to vote this one down, though actually there are some mildly interesting issues once you get past the silly (and flawed) reflection example.
Voters For This Link (9)
Voters Against This Link (6)