RE: Test-first modelling



Date view Thread view Subject view Author view Attachment view

From: Jason Gorman (jasong@xml-objects.com)
Date: Sat 13 Jul 2002 - 13:36:27 BST


Interestingly enough a lot of people point to model refactoring as a major stumbling block to an "extreme" approach. Agile Modelling seems to recommend throwing away diagrams when they're out of date, which is one approach, I suppose.

It occurs to me that there are two things that would help enormously:

1. The ability to capture and execute model test scripts (and organise them in larger models)

2. The ability to evaluate constraints against snapshots automatically during test execution

3. The ability to refactor a model with less effort

Surely tools available for test-first programming and code refactoring in languages like Java or C# could be reapplied to some kind of executable form of UML? Could the approach work then?

Jason

----------------- Original message follows -----------------

Date:    12/07/2002 - 5:15 p.m. (UTC)
From:    "Bauer, Robert" <rbauer@rational.com>
To:      "'puml-list@cs.york.ac.uk'" <puml-list@cs.york.ac.uk>
Subject: RE: Test-first modelling

Jason,

You might wish to note that this precisely how one goes about building a
complex database; but, in order to "normalize" the database and then deal
with physical design issues for optimization, one needs the whole "thing".
The problem of course is that if you try to do build it incrementally, the
constant refactoring is worse than biting the bullet and tackling the whole
problem.

robert

-----Original Message-----
From: Jason Gorman [mailto:jasong@xml-objects.com]
Sent: Friday, July 12, 2002 5:28 AM
To: puml-list@cs.york.ac.uk
Subject: Test-first modelling


Hi

I'm curious to know if anyone out there has been experimenting with a
test-first approach to building models (?) I know many analysts and
designers flesh out models using use case scenarios to discover needed
operations, but has anyone gone further and managed to get the process down
to a finer level of granularity? Does anyone build and test their models one
scenario of one operation at a time? Does anyone use "mock types" - the
modelling equivalent of mock objects used by many XP-ers? What about
regression testing of models, or model refactoring?

Regards

Jason Gorman
Independant Consultant
jasong@xml-objects.com
+44 7785 712367



To remove yourself from this list please mail puml-list-request@cs.york.ac.uk
with a message containing the word "unsubscribe".

Date view Thread view Subject view Author view Attachment view