1. ## Cause Effect Graphing

I've read some literature on Cause Effect Graphing as a means to produce test requirements but I have a question that hopefully someone can answer. Basically, can a combination of causes trigger more than one effect or is there only ever one effect for a set of combinations?

2. ## Re: Cause Effect Graphing

The more I thought about this, the more tricky it seemed. When I have done Cause-Effect graphs, I designed a set of cause values that will make an particular effect manifest itself (or not). I have implicitly assumed that other effects may manifest themselve but these are not of interest to me so I could ignore them.
It is like a physics experiment where the aim was to keep everything constant except for the causes and see if the results confirm the theory.
Paul

3. ## Re: Cause Effect Graphing

"Basically, can a combination of causes trigger more than one effect or is there only ever one effect for a set of combinations?"

Paul is correct. If the set of causes is *identical each time, the effect will remain constant. It is possible for the effect to manifest itself in several locations. Not really what you wanted to hear.... Concentrate on the function under test, if you happen across multiple effects, document all noticed effects. Just remember - QA reduces risk, it doesn't ensure bug free software!
-Mark

4. ## Re: Cause Effect Graphing

Thanks for the feedback,
I agree with Paul in terms of some effects are not of interest to the test. The situation I had in mind is as follows. Say you are creating a new customer record and the Title, First Name and Surname are mandatory (other fields are optional). Leaving any one of these blank causes one of three error messages to be displayed. These can be modelled as the effects. When you leave any combination of 2 of the mandatory fields out you get two error message displayed and if you leave all three blank then you get an error message for all three.
Concentrating on just the validation of the error messages, you might have 3 effects (one for each error message), combinations of inputs where one field is missing causes 1 effect. A combination of any two missing fields causes 2 effects to be triggered (both of which you are interested in). Leaving all three blank causes all 3 effects to be triggered.
If you are interested in validating the correct error messages are being displayed and each cause set can only relate to 1 effect then you would need to have one effect for each combination of the error messages (minus any redundancies).

Would you model situation as 1 CEG (as described above) or as a 2-3 CEGs e.g. one of single errors, one for combinations of 2 messages and one for 3 messages (probably not requrired!)?

All the examples of CEG that I've seen only have a single effect for a given cause set, but that said, the examples have been either fairly simple or described in such a way that only one effect is possible for each cause set. These have been academic papers so i'm interested in understanding how people use CEGs in the real-world.

Many thanks
Regards
Bill Matthews

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•

vBulletin Optimisation provided by vB Optimise v2.6.0 Beta 4 (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.