Re: [xtm-wg] TM Conceptual Model: Semantics of the UML "modern dinosaur"?

[Murray Altheim <murray.altheim@eng.sun.com>]

"W. Eliot Kimber" wrote:
> 
> Murray Altheim wrote:
> 
> >   using UML itself to clearly and concisely create the model. IOW, if
> >   the UML doesn't suffice to describe the model, why not just use text?
> 
> This is an important statement. There's no particular magic in the UML,
> it's just a tool for narrowing the focus of a particular discussion and,
> hopefully, to provide an effective way to distinguish, for example,
> abstract ideas (data model) from implementation ideas (representation
> syntax). This is, in essence, no different from using a DTD to define
> document structure constraints--at the end of the day, what's critical
> is the words, not the formal syntax, but the formal syntax is still
> useful and important.

Certainly.
 
> But symbolic formalisms are always valuable--at a minimum, they provide
> a shorthand that speeds communication and understanding.
> 
> Please do not throw out this particular form of formalism just because
> the supporting prose is missing or inadequate--rather, provide the
> necessary supporting prose.

Prose and diagrams. A legend. This is essentially what I was saying; that
absent the necessary explanations, the UML is currently definition-less.
But I'm still uncertain whether it has enough 'teeth' to be truly formal.
I look at the current diagrams and don't see a formalism. If I didn't 
see all those textual notes, I begin to believe it was a formalism.

> By the same token, if you can't summarize the model clearly in a few
> paragraphs, there's probably a problem either with the model or with the
> collective understanding of the model, either of which would need to be
> fixed.
> 
> The primary reason I like using UML (or any similar graphical modeling
> language) to do modeling is because it makes it easy for me to draw a
> concise picture that I can then point at in order to say "this is the
> scope of what I'm talking about now", e.g., "I'm talking about the
> abstract data model" or "I'm talking about the implementation model" or
> "I'm talking about somewhat tangential metaphysical stuff". Without a
> symbolic formalism that makes these domains more distinct, it can be
> difficult to either convey the distinction in order to establish
> conversation scope or for a newcomer to quickly intuit scope.

I completely understand. When you and I are sitting with a beer, one
of us will naturally reach for a pen and a napkin and begin with a 
diagram. It's what the first humans did in the sand.

> How many conversations have we collectively had that ultimately ended
> with "Oh, I was in the X domain, not the Y domain. Nevermind". The
> number of these can be reduced if you have clear ways of distinguishing
> the domains. We have demonstrated that prose-only standards seldom, if
> ever, do a good job of this, simply because prose does not provide many
> tools for making such distinctions clearly and the cost of using the
> tools it does provide well is high.

As I mentioned in my previous message, I can certainly see the value in
using UML internally as a tool for discussion. I'm just less convinced
that we can gain the same level of understanding, and not actually add
to the confusion, by promoting our flavour of UML as a formalism. And
yes, if the UML ends up in the final spec (which it looks like it will),
then I think we need to provide concise descriptions of exactly how we're
using it. This is different than BNF, in that there can be many different
interpretations of a box and an arrow. I'd also like to minimize the use
of the textual notes, as they don't improve the diagrams themselves, and
should simply exist in the prose of the Conceptual Model section. 

[And as I hope I've already expressed to you privately, I can only thank
you for the trouble you've taken in exposing us to UML and our use of
it, and in writing that tutorial. I do think this has been valuable work.]

Murray

...........................................................................
Murray Altheim, SGML/XML Grease Monkey     <mailto:altheim&#64;eng.sun.com>
XML Technology Center
Sun Microsystems, 1601 Willow Rd., MS UMPK17-102, Menlo Park, CA 94025

      In the evening
      The rice leaves in the garden
      Rustle in the autumn wind
      That blows through my reed hut.  -- Minamoto no Tsunenobu

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com