The Model Driven Software Network
Raise your level of abstraction
Are there any attempts to get the same MDSD approaches into unit testing?
Hi,
some time ago I have implemented unit tests with NUnit to apply unit testing for an enterprise service based project. At that time I have written the tests manually.
Now I have several more books and also one documenting xUnit tests as patterns, thus there is a more or less point where tests are similar. This is the nature of patterns.
Because of this and my involvement in code generation and modeling tasks I ask myself if there are any attempts of creating typical tests based on the occurrence of typical modeling artifacts.
Are there any attempts or documents available on the net?
Thanks
Lothar
some time ago I have implemented unit tests with NUnit to apply unit testing for an enterprise service based project. At that time I have written the tests manually.
Now I have several more books and also one documenting xUnit tests as patterns, thus there is a more or less point where tests are similar. This is the nature of patterns.
Because of this and my involvement in code generation and modeling tasks I ask myself if there are any attempts of creating typical tests based on the occurrence of typical modeling artifacts.
Are there any attempts or documents available on the net?
Thanks
Lothar
Views: 0
Replies to This Discussion
-
Permalink Reply by Gordon Morrison on -
Lothar,
I've written a book on how to improve the quality of software thereby making testing more about the correctness of the logic and data. I've learned that testing does not improve the quality of software therefore testers have truly thankless jobs that are "critical" due to the current nature of software. Removing all of the bugs from poor quality software still leaves poor quality software.
I’ve been selected as a presenter at the 2010 Systems & Software Technology Conference in Salt Lake City. I’m excited about being selected and looking forward to presenting.
The theme of the conference is "Technology: Changing the Game".
From the conference website: “This year, the Systems and Software Technology Conference will explore various technologies which are expected to make abrupt changes to common thought. We will explore the tools, the processes, and the ideas which will "change the game" and make the way we have done things in the past - obsolete.”
Regards,
Gordon Morrison
Author: Breaking the Time Barrier-The Temporal Engineering of Software
Inventor: Multi-Core & Hyper-Threading Technology (US Patent 4,847,755)
Senior Member IEEE; www.vsmerlot.com
-
Permalink Reply by Mark Dalgarno on
-
It's of little help to you other than as a motivating example since there's no public documentation on it but we were quite successful in my previous company at generating unit tests for the COM Automation and OLD DB interfaces into our products.
-
Permalink Reply by Karsten Thoms on -
Hi Lothar,
MDSD has always 2 sides: What can be and is done in projects, and what is available for public. So, as Mark, I just can give you the answers that generating unit tests of course is possible and practice.
On the W-JAX in Munich next week I will talk about my project at the Deutsche Boerse, where we also generate unit tests and also integration tests to some extend. The created tests are very domain specific, and they are just a rather small part of what is done for testing in this huge project.
~Karsten
-
Permalink Reply by Rui Curado on -
"I ask myself if there are any attempts of creating typical tests based on the occurrence of typical modeling artifacts."
As you present it, I guess we could consider unit testing as a cross-cutting concern. However, due to the complexity involved, I am not sure if an AOP methodology could cover that.
Currently I am developing AtomWeaver using the ABSE model-driven methodology, and I have implemented unit tests using models. However I implemented tests as models themselves, and not as a cross-cutting concern as you present it.
© 2012 Created by Mark Dalgarno.
Powered by
