[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: (FC-Devel) Implementing FreeCASE

Being new on this list, I suppose I ought to introduce myself first: I'm a
33 year old perpetual computer science student at the univ. of Skövde,
Sweden.  I'm very glad I found this project as I was just about to go
through the tedious process of starting up a similar project myself.  My
original plan was to rewrite KTG's ood to UML and freshen it up with a
better toolkit + using C++'s STL.  After taking a closer look at the source
for ood, I think it's safer to view it as a prototype, and write something
new, perhaps snatching a few pieces from it.  (ood isn't "bad"--it's just
that many things has happened since ktg started working on it)
I'm not a software-engineer; I only know enough about it to realise just
how little I know. =)  (I'm trying to improve by plowing through

And now for something completely different:

On Wed, 21 Apr 1999, Sami BENGRINE wrote:

> So I have sbeen looking for a platform for implementing a Core FreeCASE 
> program (which means able to implement the static models of UML with a GUI).
>  I am currently working with C++ because I feel more comfortable with it 
> than Java (yes, what about Argo then ?) and because Qt (http://www.troll.no)
>  seems to be a nice GUI library. I have quite a clear idea on the client 
> architecture (with MVC).

(Sorry, "MVC" ?:-P)

I like this plan too. (i.e. C++ and Qt)  The Java-people already have a
couple of projects in java; I'm specifically interested in C++, both for
implementing the tool itself, as well as code-generation and
reverse-engineering.  ...I don't have anything /against/ java, I just like
C++. ;)  I'm also specifically interested in developing on/for unix

The way I see it, the project could be viewed as consisting of a few main
parts, that at least initially could be developed separately:
1. A frontend (the user interface)
2. A repository (there are work in progress on this I believe?)
3. A reverse-engineering unit (this can be a /lot/ of work I think --
	someone with good skills in compiler-construction ought to start looking
	into it)
4. Code-generation (performed by each `FC-object' visualised in the frontend)
and perhaps a
5. Metrics generator  (this could be completely separate from the others;
	there are also free metrics report generators available...well, at least

I would like to start working with the frontend right away [well, that's
the easiest part =), but it also good for the moral to have a prototype,
something to try out ideas about the user interface with]

Does anyone have any code for a frontend?  A stub, a test, anything?

I only have a very rudimentary test-app lying around: just a QMainWindow
with a QScrollView, plus a widget-class that can be moved around in the
scrollview and have its own popup menu. ...only made it for experimenting
with Qt when I was still on to porting OODesigner to Qt, so it's not meant
as a prototype. Unless there is something better around, I could start
cleaning it up a bit. (Hmm...it doesn't even compile at the moment, I
started adding something to it last time I touched it, now I forgot what)

The meek are inheriting Earth.