software engineering question

Toggle sidebar Toggle sidebar
C

canyouhelpme

Member
Jun 27, 2001
69
0
0
Have this question in my final exam

?When generating test plans you should select tests based on the likelihood that they will find faults. The testing group should be evaluated on how many errors they find?

Anyone here know how to prove that this statement is valid?

Thanks in advance.
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0


<< ?When generating test plans you should select tests based on the likelihood that they will find faults. The testing group should be evaluated on how many errors they find? >>


This part sounds like BS to me.

But then I never took a software engineering course at college before. I follow MSF and we rate bugs based on priority and severity. Otherwise, the team does risk management on a regular basis.
 
her209

her209

No Lifer
Oct 11, 2000
56,352
11
0
I assume a Test Plan is a plan for the the testing procedures.

Test cases are specific tests specified in the Test Plan.

Its been almost 2 years since I took my Senior Project class tho.
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0
I should know, but I don't. :eek:

I haven't played the testing role yet.

But I think that her209 is correct.
 
C

canyouhelpme

Member
Jun 27, 2001
69
0
0
When generating test plans you should select tests based on the likelihood that they will find faults

I know if "When generating [test cases] you should select tests based on the likelihood that they will find faults"

tat statement will be valid...

but what about test plan? is it the same?
 
A

Ameesh

Lifer
Apr 3, 2001
23,686
0
0


<< Have this question in my final exam

?When generating test plans you should select tests based on the likelihood that they will find faults. The testing group should be evaluated on how many errors they find?

Anyone here know how to prove that this statement is valid?

Thanks in advance. >>



I dont think the statement is valid at all,

when making a test plan your objective should be to make the most amount of coverage against the compnent testing for correctness in the least amount of tests. You are looking to make sure the code does what it is supposed to do in every scenario. you have no idea which tests are gonna find how many bugs thats why you have the tests. The best thing you could do is make sure you cover all aspects and functionality of the code itself so you get a better idea if your are finding the bugs that will affect consumers.
 
C

canyouhelpme

Member
Jun 27, 2001
69
0
0
Thanks Joo Hang

When generating test plans you should select tests based on the likelihood that they will find faults.

does it sound valid to you?
I am desperate here...i do not know where to look for the correct answer.
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0


<< Thanks Joo Hang

When generating test plans you should select tests based on the likelihood that they will find faults.

does it sound valid to you?
I am desperate here...i do not know where to look for the correct answer. >>


OK. I'll grab my MSF binder and see what I can find out.

But don't take my words 100%, though. I have almost no experience in testing. The closest thing I did is buddy development.
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0


<< I think its ironic that Microsoft makes that MSF program. >>


LOL.

I heard that from many other people. :)

MSF is not very well-known within Microsoft, but it developed within Microsoft over the years. A lot of core elements of MSF are adopted in many departments within Microsoft, though. However, many Microsoft employees didn't even hear about such thing as "MSF" and the MSF department is quite small. It is no wonder that Ameesh once asked me what MSF stood for. :)

MSF is pretty much what killed Borland off the compiler market and made Visual C++ and Excel 3.0 take up so much market share over their competitors. One of the "founding fathers" of MSF is Jim McCarthy. The company I work for is closely endorsed with Jim's company.

Here's a link to Jim's most recent book, if you are interested. It is one of those great books that every IT people should read.
Linkie to Amazon
Book Review
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0
Sorry. Didn't find much stuff on my binder, but I could ask.

Actually, I think I'll just ask any ways because I'm curious to know.
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0
Here's the response from my boss and software architect. This is a response from a certified MSF trainer.



<< Test Plans are the hows and whens of the testing aspect of a project. The test cases are individual tests that are run against the system.

Testers are responsible for ensuring that the system meets the requirements, and that the implementation of those requirements are above the quality bar. A tester is not evaluated by the number of bugs that she finds. A tester is evaluated the same way the rest of the team is: whether the product ships, and how good it is.

Writing your test cases first refers to regressions tests. It is a good coding practice to write your automated regression test before you write your code. This ensures that you will know when your code works. >>

 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0
From an MSF standpoint, it is not. Or at least MSF approaches software development in a different perspective.
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0


<< How should the testing group be evaluated? >>


Will you define what exactly you mean by "evaluated"?
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0


<< I am not sure either, i copied it out straight from my final paper. >>


LOL. :)

If this is some way to measure the "performance" of the testers as a whole, we don't do such thing with MSF.

In MSF, the entire team is evaluated as a whole. Since all team members have rich interactions with other members across their roles, it is contradictory to evaluate the testing role as a separate entity. Not to mention that testers are developers in their own rights, and sometimes developers and testers switch roles in MSF.
 
C

canyouhelpme

Member
Jun 27, 2001
69
0
0
found this from this website

when making a test plan your objective should be to make the most amount of coverage against the compnent testing for correctness in the least amount of tests. You are looking to make sure the code does what it is supposed to do in every scenario. you have no idea which tests are gonna find how many bugs thats why you have the tests. The best thing you could do is make sure you cover all aspects and functionality of the code itself so you get a better idea if your are finding the bugs that will affect consumers.

Text

Just wondering whether there is no correct answer for this question?
 
J

joohang

Lifer
Oct 22, 2000
12,340
1
0
There are many answers.

Oh yeah, in MSF, testing also varies depending on team Design Goals.
i.e. Availability, scalability, usability, reusability, extensibility, maintainability, etc.

We define 3 primary design goals and up to 3 secondary design goals.

So the testing process should reflect the requirements of these design goals. For example, if maintainability is one of the top design goals, we do more frequent code reviews and make sure that it is readable, follows the industry Best Practices/naming conventions, commented appropriately, etc.
 
C

canyouhelpme

Member
Jun 27, 2001
69
0
0
How about this question?

List the characteristics of a system for which prototyping is most appropriate.


Any ideas?
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom