----- Original Message -----
From: Rich Edwards <[EMAIL PROTECTED]>
To: Robert Simmons <[EMAIL PROTECTED]>
Sent: Tuesday, October 12, 1999 6:44 AM
Subject: Re: Jump and ArgoUML
Prior to responding to you. a little poll for everyone. For the interface,
would you all prefer a desktop interface with internal frames or a tab pane
based interface. Since im working on infrastructure right now this is
totally changable right now.
> Hi Robert --
>
> I just caught the end of this discussion, and didn't see any links to
> more info on what you're developing (perhaps there aren't any yet).
> I, like you, am a professional software engineer who hasn't had much
> value for the drawing-based design tools -- too much effort to learn,
> not enough payoff. I'm very interested by this project, as it seems
> like a way to make UML more of a real-world design tool.
>
I know that feeling. =) UML is a conceptual tool that should provide all the
information on a project that is needed to understand the project, not just
a pretty drawing for inclusion into documentation. Further, it should have a
very shor learning curve. Im even debating including a power point
presentation tutorial on how to use it with the documentation.
> We have been using UML to an extent on our latest project, but only as
> a means of standardizing on symbols and concepts. However, what we
> have found is that UML doesn't do a very good job at modeling
> distributed designs (i.e. objects defined in OMG-IDL, and living in 15
> to 20 separate processes). Have you considered how to handle such
> distributed designs with your tool? To me, that is what most of the
> other tools are really lacking -- they seem geared towards the
> its-all-one-big-program design model.
Actually the focus in my Software engineering career is on CORBA based
systems so you can be sure that I shall address those issues. Keep in mind
that what I have in in the early stages. However it is progressing rapidly
and I am counting on the thousands of developers out there to give it even
more punch. Farfetched ? Yeah, thats what they said about emacs. You will
notice that if you run make doc for the stuff I packaged up with the
origination of this thread that you get FULL documentation. Novel idea.
Contrast that with argos .. well I guess you could call it documentation,
take a look at it.
>
> It's easy enough to describe an object-oriented design with simple
> class diagrams, but where the rubber meets the road is when you begin
> to design where the various objects will be deployed (e.g. in what
> process, and on what machine). The UML would have you believe this
> isn't really an issue (there isn't even a concept of a 'process', that
> I'm aware of, and the so-called 'component' is a cumbersome,
> non-intuitive symbol that is virtually unusable). Those of us
> designing real systems know that object deployment is one of the key
> design decisions that shapes a distributed design.
Actually UML's backers do not believe this is a non-issue. Whomever told you
that didnt really give you good information. There are a number of
strategies for modeling CORBA based systems (ie: interface - implementation
split) systems. Also there are a couple ways of modelling runtime structure
in UML. But then again, I actually give seminars on UML so perhaps I know a
bit more about that then some people. There are alot of people out there
that would have you believe UML is just for class diagrams in a monolithic
language. Not so. Certainly my project will take that into account. Further,
it will be designed for use within a professional environment, with things
like source control considered.
>
> Anyway, I'd like to hear more about your project. The extensibility
> you're building into it sounds promising, and the fact that you're
> going to be using the GPL makes it even more attractive.
Yepper! Id rather be famous then rich =) I will be using it to instruct
students of my seminars as well. Im tired of making them swim through
Rational Rose which is unneccesarily difficult to use. My proudest moment
will eb when the first person extends it. Remember KISS (Keep it Simple
Stupid) =) I want it to be extensible and easy to understand. As a matter of
fact for people that want to make their extension stick in the official
package, If their extension is documented like argo, it wont get in the
official release. I WANT people to understand the structure.
>
> Thanks,
> Rich Edwards
>
> --
> Rich Edwards
> Senior Software Engineer
> Codonics, Inc.
> email: [EMAIL PROTECTED]
> web: www.codonics.com
> voice: 330.665.1210
> fax: 330.665.1169
> -- __o
> _ \<,_ "I ride my bicycle to ride
> (_)/ (_) my bicycle." - Zen proverb
>
>
> Robert Simmons wrote:
> >
> > *shrug* I suppose there is somewhat of a duplication. Im not sure of the
> > license of argo. My project will be released with full open source.
> > Personalyl, knowing universities, I dont trust them to come out with
> > comething completely open. It could turn out that the project is largely
for
> > my own good. It is somewhat deflating to find that there is already
another
> > project out there doing what Im working on but competition is supposed
to be
> > a good thing. I can tell you already that my project has a very
different
> > focus. It is NOT a glorified drawing program. As a matter of fact, other
> > then the placement of objects, I am not planniong on implementing
drawing
> > features at all. I have always believed that modeling tools were to
heavy on
> > drawing and not heavy enought on UML modelling, which is largely
concerned
> > with classifiers and their relationships to each other. I have never
> > believed that a UML tool should have a focus on art. In fact many tools,
> > Visio, Rational Rose, have so much emphaisis on drawing that the
features of
> > UML are BARELY enabled within them. Other tools are highly language
> > specific, with many generic features of UML Missing. I am a senior
software
> > engineer professionally. I have found that most university based people
have
> > limited real life knowledge of professional software engineering. My
tool
> > will be disigned with the professional.
> >
> > I have glanced at the argo web page. First thing that occurs to me. THe
> > documentation is barely there! No javadoc comments, just a laundry list
of
> > mehtods. No model to the program. You would think the software would
have a
> > model of at least itself. I know Jump will when released. Not very
conducive
> > to contribution. It looks very java centric, not general. It also looks
more
> > graphical then modelling.
> >
> > --rob
> >
> > ----- Original Message -----
> > From: Sean Chen <[EMAIL PROTECTED]>
> > To: <[EMAIL PROTECTED]>
> > Sent: Monday, October 11, 1999 1:22 PM
> > Subject: Jump and ArgoUML
> >
> > > Hi Robert,
> > >
> > > Was just made aware of your project from the java-linux list.
> > >
> > > Are you familiar with ArgoUML http://www.ics.uci.edu/pub/arch/uml/
> > > at all? Would it be possible to integrate your work within it
somehow?
> > >
> > > I just hate to see duplication of effort =)
> > >
> > > By the way, your makefile has a bug: if Jump.jar doesn't exist then
jar
> > > -umf will fail of course.
> > >
> > > . . . Sean.
> > >
> > >
> > >
> > >
> >
> > ----------------------------------------------------------------------
> > To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> > with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
