Wednesday, June 25. 2008Using the Overview PyramidTrackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Firstly, thanks for undercontrol and depend.
Very interesting, these new feature. just a question, on what are base the reference values ? how are they determined?
Hello Davide,
this set of reference values was taken from the book *Object-Oriented Metrics in Practice*. The book mentions that these values are based on 45 Java projects (a mix of commercial and open source applications) of various size (20.000 up to 2.000.000 lines of code).
Hi!
Again: Thanks for pdepend. It seems to be kind of comprehensive work. In fact I bought Object-Oriented Metrics in Practice some weeks ago, before trying out pdepend. It is a cumulative work. Now I am pretty happy, that you reference this one. Cheers Jan
"""In a system of ten classes an ANDC-value of 0.5 means, that every second class is derived from another class."""
I reckon it would be every second class, no matter how many classes we talk about?
That's true, 0.5 means always the same, independent of the class count.
I love metrics, but metrics for the sake of coming up with numbers which I must say mean have no objectivity and no measure of usability in practice is just benchmarking for the sake of benchmarking.
And I must object to the use of Java baseline figures as a a basis for PHP code metrics. A single line in one language may translate to several in another, and vice versa. So that distorts your figures and makes it totally fallible and unreliable.
Any specific reason why this information is presented in the quirky "pyramid" format? It seems that three columns of a standard spreadsheet would suffice (label, number, ratio). Using a standard presentation scheme would make the information more intuitive to read.
Yes, there is a reason. The three corners of pyramid represent three different metric-categories:
Size and complexity - left Coupling - right * Inheritance - top |
ProjectsFurther stuffCategories |

