RE: [ihc] UML Tools

[Rex Brooks <rexb@starbourne.com>]

I misspoke myself, Sylvia,

Thanks for replying so quickly since it gives me a chance to clear 
this up by copying the ihc list.

I did indeed mean what you said--that it does not support use case, 
sequence or activity diagrams, which is the way I think of UML 
modeling. Basically, the way I work is to model the process, 
especially if there are existing standards which I want to 
incorporate that do not have supporting models and where I can't 
reverse engineer an application based on the standard to extract a 
model from which to determine if it is complete enough in itself for 
a given set of requirements.

When I model the process, if there is no standard for a particular 
area, I generally look at existing applications if I can or I develop 
a strawman vocabulary from a problem statement or a business scenario 
that describes the processes which need a standard or for which an 
interest group has decided to create a standard.

The models which standards bodies need to create most are use-cases. 
Use cases are usually developed from a business or process scenario 
and/or a problem statement that identifies elements of a vocabulary, 
such as actors and categories of functionalities such as message 
types between actors which is used to make up the use case.

It is then from use-cases that requirements for a standard are drawn, 
and a group starts working on definitions for vocabulary terms, 
message types, actions or processes, etc.

In the IHC TC we are still at the survey stage, getting a landscape 
view of the problem space we are in the process of scoping. We are 
identifying areas of functionality and stakeholders from those areas. 
 From there we will probably divide up the work according to 
categories that we think belong together or share a particular kind 
of functionality such as medical records, physician credentials and 
insurance-financial-legal concerns. From there a more intensive 
survey should yield scenarios from which use-cases can be drawn 
requirements established.

One thing I forgot to mention about both Visual Paradigm and GEFEG is 
that there is somewhat more allowance for the need to create 
documentation than I have seen in such tools previously, and this 
addresses one of our major concerns to an extent.

What EDIFIX offers that is most valuable in my estimation is the 
ability to generate an xml schema from the models we prepare and then 
to generate testable classes.  To be honest, I have not yet had an 
opportunity to approach a standards assessment this way before, so i 
will discover if this process is the logical extension of the 
processes I have worked with previously. For the vendors in the TC I 
think it will prove extremely valuable in saving their production 
teams a great deal of time developing components from which to build, 
replace or enhance applications.

Ciao,
Rex


At 8:26 AM -0800 12/27/04, Sylvia Webb wrote:
>Hi Rex,
>
>Thanks for the update.
>
>I'm a little confused.  EDIFIX does allow you to create UML data models and
>CCTS data models.  In fact, I'm working on a manual for creating CCTS data
>models now.  EDIFIX doesn't have the pretty GUI that other modelers like
>MagicDraw, EA Enterprise form Sparx Systems and others have but it does
>allow you to create fully compliant UML models. From those models, you can
>create, schema, sample schema, analysis maps, and documentation.
>
>EDIFIX does not support the creation of use case diagrams, sequence, or
>activity diagrams.  If this is what you mean by the statement that it is not
>a modeler, you are correct. 
>
>Can you help me understand more about the type of model you need to create?
>
>Thanks,
>Sylvia
>-----Original Message-----
>From: Rex Brooks [mailto:rexb@starbourne.com]
>Sent: Monday, December 27, 2004 8:37 AM
>To: ihc@lists.oasis-open.org
>Cc: swebb@gefeg.com
>Subject: [ihc] UML Tools
>
>Hi Everyone,
>
>I wanted to give you all a brief update about what I have learned about both
>GEFEG's UML product EDIFIX and a free UML Modeler that I may be using
>until/if we are allowed a reasonable legnth of time for an evaluation or
>not-for-profit license for IBM's revamped Rational toolset when it is
>released sometime in January.  That UML modeler is available from Visual
>Paradigm, under its "Community Edition" at
>
>http://www.visual-paradigm.com
>
>Visual Paradigm is a Hong Kong company, so the tutorials have some
>interesting translation constructions, but it makes for a nice International
>Package with the German GEFEG's EDIFIX product.
>
>Just FYI, Gefeg's EDIFIX is not a modeler. What it does is take a UML model
>and output an XML Schema and documentation as well as basic code from the
>model which can then be used by our various member companies to test how our
>candidate specifications perform, or how our thesari or equivalence-mapped
>vocabularies perform. Likewise, VP for UML is a full featured UML modeler,
>and it also outputs or reverse engineers program code which can also be
>output to work with the major IDEs, with a separate modeule for Visual
>Studio .NET using Visual Paradigm Suite which includes Smart Development
>Envvronment for Eclipse, NetBeans, JBuilderIntelliJ IDEA and WebLogic
>Workshop.
>No, I don't own stock in this company. I am slogging through the tutorials
>along with the EDIFIX Manual..
>
>As far as I can tell at this point, just about the only things that this set
>of programs lacks that Rational had is the ability to track project
>development, along with runtime simulation testing and monitoring, which was
>one of the key features of Rational that I was looking forward to using.
>However, who knows what will be available in the new Rational suite that is
>being developed under the code term Atlantic. Hopefully, IBM will release
>the newly reconfigured Rational Suite at about the time I will be ready to
>test it with their Websphere Application Server v6.0 which is now available.
>Since I work cross platform and maintain a vendor agnostic approach to
>Enterprise-wide IT Integration, I just happen to be very interested in these
>intersections. They also tend to make the standards we are investigating
>somewhat more universal in applicability.
>
>Regardless, I just wanted to make sure you all know that we have relatively
>full-featured toolsets available on which to build and test our standards.
>
>Ciao,
>Rex
>--
>Rex Brooks
>President, CEO
>Starbourne Communications Design
>GeoAddress: 1361-A Addison, Berkeley, CA, 94702 USA, Earth
>W3Address: http://www.starbourne.com
>Email: rexb@starbourne.com
>Tel: 510-849-2309


-- 
Rex Brooks
President, CEO
Starbourne Communications Design
GeoAddress: 1361-A Addison, Berkeley, CA, 94702 USA, Earth
W3Address: http://www.starbourne.com
Email: rexb@starbourne.com
Tel: 510-849-2309