On Mon, 2007-07-02 at 15:51 +0200, Tudor Girba wrote:

Excellent. I added it to the wiki.

I noticed you use Measurement.

We used to have Measurement as well, but this proved to induce a lot of overhead because we ended up with tons of Measurement objects in memory. As a consequence in our current FAMIX, we store measurements/ properties as attributes directly in the entity Class. For example, for number of methods, we would have a FAMIX.Class.NOM attribute, while you would have a Measurement object.

Would it be a problem for you if you would adapt your tools to not have explicit entities for Measurements?

Let me first clarify this design decision: - As we are working with tuples in Crocopath, changing the representation as well as adding additional measurements in the future (which is likely to happen) requires changing all our querying scripts every time. - Secondly, some measurements (Coupling & Cohesion) are calculated as a post processing step. Being a seperate entity, it is easy to append them to the fact base. This means that it was pure convenience to implement it like that.

However, there is a difference in what constitutes our meta-model and how we put in in a certain file format (read RSF, or, MSE). The presented meta-model is extracted from our RSF representation. Producing MSE with 'inline' measurements is not a problem for us. Therefore, if you want to make it an attribute of the respective entities that's fine for us.

Another option that you might consider is gathering measurements per entity in a QuantitativeData entity - which would only double the amount of objects in the system at most.

cheers, Bart

Cheers, Doru

On Jul 2, 2007, at 10:29 AM, Bart Van Rompaey wrote:

...

On Sun, 2007-07-01 at 19:35 +0200, Tudor Girba wrote:

...

Hi,

LOOSE Research Group will also participate in the definition of FAMIX 3.0.

Radu, could you give a short overview of MEMORIA? I would say a UML diagram would do just fine.

The input for the discussion up to this point is given by:

• SCG FAMIX:

http://smallwiki.unibe.ch/moose/famixmetamodel/

• SEAL FAMIX:

http://seal.ifi.unizh.ch/wiki/swiki/show/FamixCore

The UA likes to contribute their meta-model as well:

http://www.win.ua.ac.be/~bdubois/images/metamodel.png

Its a RSF-alike model transformed from FAMIX CDIF.

cheers, Bart

...

Cheers, Doru

-- www.iam.unibe.ch/~girba www.iam.unibe.ch/~girba/blog/

"Every thing has its own flow."

---

Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- Bart Van Rompaey Lab On Reengineering Dept. of Mathematics and Computer Science University of Antwerp Office location: CMI - G.2.22 Office phone: +32 3 265 34 05 URL: http://www.win.ua.ac.be/~bvromp

-- www.iam.unibe.ch/~girba www.iam.unibe.ch/~girba/blog/

"No matter how many recipes we'll know, we'll still value a chef."