Software metrics for OO programs

Software metrics for OO programs

[Chandrashekar Rajaraman]

Hi,

I need this stuff urgently.

Can you name some research papers or books
that provide some information about
size metrics for OO programs? Are there
any metrics out there other than SLOC to 
measure OO program size ?

How about complexity metrics for OO programs ?

How about evaluating a design for reusability ? 
Any metrics out there for it? 

Are there any papers/studies about the application 
of Function Points to OO programs ?

Please send me email replies. If you want to
post it on the net, then please do so on comp.software-eng.

I will post a summary on comp.software-eng, if 
there is good response.

Thanks a lot for all your help, in advance.

I also need some help on an old posting by Sharon Crichton.
I was off the net for a while, so I would have missed
a summary of the responses that she may have posted.
Can someone out there please email me her summary(if any) ?
Can you, Sharon, please mail me all replies that you
received (if no summary was posted on the net) ?
I enclose her posting below.

{
Hello to everyone!

I'm currently trying to solve a problem/situation that may occur with Object-
Oriented Analysis/Design/Programming (insert your favorite OO acronym here).
We are investigating using OO for a project, and we have some specific goals
that we want to achieve by using OO design:

1. Encapsulation - standard OO encapsulation advantages

2. Flexibility and Extensibility - the ability to accomodate future scenarios
of potential new features with no or minimal changes to the design

3. Ease of Implementation and Maintenance - the ability to design classes in
such a way that all classes are (relatively) easy to code, with a minimal amoun
t
of control (e.g. IF statements, switches, flags, etc.) embedded.

4. Reusability - standard OO reusability of classes advantages

I know that I can start a lot of flame wars about the middle two, but that's no
t
my basic question. My question concerns whether you can reliably test and
measure how well your design or code has reached these goals. For instance,
how do you measure how well a class is reusable: by how many other classes
uses it, or by how many classes are repeating code that can be abstracted into
a reusable, inheritable class? I know that the middle two are going to be hard
to measure in any testing methodology.

Are there any good texts out there that specifically address the concept of
measuring OO design/code to meet OO goals? Or does anyone out there in net.land
have any experience in this area? We really would like to get as much input as
possible to help guide us along this rocky path.

Replies via e-mail preferred. I only read comp.lang.c++ on a regular basis, so
if you wish to post a reply, please do so there. If there is enough interest,
I will post a summary of replies. I've tried to hit all of the groups that may
have interest in object-oriented issues.

Thanks in advance for any and all replies!

-- 
Sharon Crichton                          CE-SAS
sharonc%meaddata@uunet.uu.net            Mead Data Central
sharonc@meaddata.com                     P.O. Box 933
uunet!meaddata!sharonc                   Dayton, OH 45401

}

Thanks again.

Chandra