I am preparing for the ISEB Practitioner Cert in Test Analysis next week, through self-study (lost my job recently so can’t afford a course), and I’m struggling a bit with equivalence partitioning / boundary value analysis so I’d appreciate if anybody familiar with the syllabus / exam could give me some pointers.
My main problem is that the ISEB syllabus is somewhat sparse saying only generally that the points to cover for EP/BVA are multiple dependent or independent variables (i.e. does not actually describe any analysis techniques). It references Beziers Black Box Testing, which I have but find it pretty heavy going – its very theoretical. I also have Copeland's Practitioner’s Guide (also on the reading list) which helps but I still have some areas I’m not sure about, e.g;
* How to distinguish dependent / independent variables – for example for a graphical display scenario, where the input variables are the x and y screen coordinates, I guess these would be independent variables, in the sense that we can choose arbitrary values for them, but I’m not sure. What would be an example of dependant variables? And more importantly, how does the methodology differ for the two cases?
* How to handle disjoint partitions (e.g x < -1000 OR x > 1000) – just treat each part separately as ‘normal’ partitions, or is it sufficient to analyse one part as proxy for all?
I have the ISEB sample paper, and attempted the EP/BVA question thusly;
1. Described the EPs in terms of the given inequalities (e.g. -800 <= x & <=0 AND 0 <= y <= 500).
2. Used arbitrary values to define test cases to cover all EPs
3. Following the above-mentioned books, performed the BV analysis by defining the ‘on’ and ‘off’ points for the lower and upper end of each boundary inequality.
4. Defined a set of test cases to ensure that all ‘on’ and ‘off’ points are exercised, avoiding duplication to produce a minimal set of TCs.
The EP part I think is fairly straightforward, but for the BVA ,I end up with a set of 28 TCs, which seems a tad excessive, especially considering this is actually only half a question, so should be doable in 15 minutes.
I am attaching my full attempted answer – maybe some kind soul could take a quick look and see if I’m doing something fundamentally silly (the answer probably doesn’t make much sense without the question, but I see this forum is pretty strict about copyright material - but maybe I could post a summary of the question?)
Also if anybody know of any good online worked examples of bva analysis for multiple (in)dependent variables I’d really appreciate it.
Apologies for the long posting, but I’m hoping this will prove to be of general interest, and not just a question of me looking for help.
Re: ISEB Practitioner Cert in Test Analysis : EP/BVA
[ QUOTE ]
* How to distinguish dependent / independent variables &#8211; for example for a graphical display scenario, where the input variables are the x and y screen coordinates, I guess these would be independent variables, in the sense that we can choose arbitrary values for them, but I&#8217;m not sure. What would be an example of dependant variables? And more importantly, how does the methodology differ for the two cases?
[/ QUOTE ]
An example of dependent variables would be a date consisting of the 3 variables Day, Month and Year. The valid values for Day depend on the values of Month and Year.
I'm planning to self-study and take the exam in June, so I'd be interested to know how you get on.
Re: ISEB Practitioner Cert in Test Analysis : EP/BVA
Thanks for the reply, Tom.
Yeah, your example makes sense. But what Im wondering then is, how to do the EP/BVA in this case (or, at least, how does ISEB expect you to approach it).
For example, would there be, rather than a single equivalence partition for day (1 <= m =< 31), a set of EPs, like {1 <= d =< 28 | 1 <= d =< 29 | 1 <= d =< 30 | 1 <= d =< 31}? And how to document the relationship between the valid EP for day and the value for month which determines it?
Yeah, I'll let you know how I get on....though starting to wonder if I shouldn't leave it til June as well [img]/images/graemlins/wink.gif[/img]