Jump to content

Talk:Software verification and validation

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Gripes

[edit]

A couple of gripes with this page:

1) V&V is not "normally part of the software testing process of a project." It's bigger than testing. Depending on a particular organization's philosophy, I've seen it as either (a) the inline-with-development structure/process that guides activities such as testing (and design reviews, etc) or (b) the parallel-to-development, classically-independent review of the output of all development phases (of which one phase is testing)

2) There is no relationship between [static testing and verification], or [dynamic testing and validation]. That is just two different types of testing; either the code is running, or not. That's just very random to me. This bizarre association is also echoed on the "Static Testing" and "Dynamic Testing" pages, except they reversed it (not that it makes any more or less sense backwards).

Also, I think I've read two sets of definitions of "Verification" with respect to "Validation". One is given in this page (is consistent with IEC-62304, others), the other says that verification is confirming the output of each phase, validation is confirming the output of the whole process. However can't cite my source; when I find it I will add to this. Oh, found it: IEEE 12207, Sections 6.4 and 6.5. (You know, on second thought, the language is so weird that both interpretations are possible. Maybe they're one and the same, it comes down to verification is low level checking along the way, validation is final check that all requirements are met. However now I wonder if "built the product right" vs "built the right product" is a handy mnemonic that loses some of the meaning -- you shouldn't wait till the "final as-built" validation to make sure you built the right product.)

199.171.110.251 19:32, 8 June 2007 (UTC)[reply]

If this article is only about....

[edit]

the software verification and validation, please add (software) in the title. Otherwise, the article will be modified for universal purposes.

I agree. A general article about verification and validation is needed. Software can have its own page with its own discussion. Rlsheehan Sept 3, 07 —Preceding unsigned comment added by Rlsheehan (talkcontribs) 00:52, 4 September 2007 (UTC)[reply]
Done. Rlsheehan 01:52, 8 September 2007 (UTC)[reply]

Building the "right product" (??)

[edit]

I am not sure that validation is normally intended to "ensure that the product actually meets the user's needs". Major software development is governed by the requirements documents, as long as the software meets the requirements then the developers have done their job and the software company has fulfilled its contract - as is described by IEEE-STD-610.

The user finds out whether the software meets their needs when they use it, but if it doesn't, then it was because they didn't agree on the right requirements, which is not really validation.

220.237.77.100 (talk) 11:13, 20 February 2008 (UTC)[reply]

See Also

[edit]

I am not sure if Atsec information security, should be in the See Also Section. —Preceding unsigned comment added by 162.84.78.55 (talk) 15:08, 30 June 2008 (UTC)[reply]

Similarly, the "see also" link to cross-validation seems inappropriate. Skbkekas (talk) 21:06, 16 April 2009 (UTC)[reply]

Disagree, here is what CBOK says...

[edit]

I think that this is confusing...

To be practical and since V & V are both used in the Quality process in the field of software development, I think we better refer to the CBOK "Common Body Of Knoledge" which is (I think) a very important reference from QAI - Quality Assurance Institute for Quality Analysts CSTA and all professional testers CSTEs.

From CBOK point of view I selected the folowing paragraph to define and show the difference between both Verification and Validation.

To cut it short: Verification answers the question, “Did we build the right system?” while validations addresses, “Did we build the system right?”.


Verification ensures that the system (software, hardware, documentation, and personnel) complies with an organization’s standards and processes, relying on review or non-executable methods.

Validation physically ensures that the system operates according to plan by executing the system functions through a series of tests that can be observed and evaluated.

Verification answers the question, “Did we build the right system?” while validations addresses, “Did we build the system right?”.

Keep in mind that verification and validation techniques can be applied to every element of the computerized system. You’ll find these techniques in publications dealing with the design and implementation of user manuals and training courses, as well as in industry publications.


Amusallami (talk) 10:15, 14 September 2009 (UTC)Ahmad Al-MusallamiAmusallami (talk) 10:15, 14 September 2009 (UTC) (AMUS) Reference: CBOK 2006 - Common Body Of Knowledge, page 113.[reply]

+++

110219 It's going to take me a bit of time to work this article up, but this is what I do for a living so I would like to think I know something about it. Also, the PE exam is coming on April 9th and I have to study for that...so, dear Review Committee, please cut me some slack and do not slate the changes I make for instant delete...instead, please contact me at xxxxxx@xxxx.xx. Thank you. —Preceding unsigned comment added by 208.96.82.76 (talk) 00:06, 21 February 2011 (UTC)[reply]

1) Please don't crew the talk page up.
2) Please reference your additions. This article already has too few.
3) Thanks for contributing.
--Walter Görlitz (talk) 04:37, 21 February 2011 (UTC)[reply]

May need some help with the formatting. Give me about a week, if you would, then jump in? Thanks.208.96.82.76 (talk) 02:22, 22 February 2011 (UTC) ...one minute later...My contribution was deleted and I have been told strongly to stay the hell away from editing anything. I use Wikipedia extensively for lookup and thought that I could pay something back concerning my particular area of specialization...apparently most definitely not wanted here, message received loud and clear. My mistake...sorry...208.96.82.76 (talk) 02:43, 22 February 2011 (UTC)[reply]

It may be the lack of references. I don't think you're contributions are unwanted, in fact, we're all warned not to bite the newcomers. If you don't respond, I'll contact you via email and discuss how to do this better. --Walter Görlitz (talk) 07:52, 22 February 2011 (UTC)[reply]

Removing the line "It is normally part of the software testing process of a project."

[edit]

The line "It is normally part of the software testing process of a project." should be removed. Not only this statement is false, it actually is exactly the opposite. The testing process is a part of the V&V process and is one of many methods that can be used. A simple example (in software v&v) is model checking which obviously is a verification method and has nothing to do with testing... Stoilkov (talk) 17:41, 11 November 2011 (UTC)[reply]

I added some of it back. You're wrong though. Testing is a also a department, just as development is, in most software development companies. That is what the lede was attempting to identify. The people in that department use verification and validation as part of their software testing activities. --Walter Görlitz (talk) 19:09, 11 November 2011 (UTC)[reply]
Many people think of testing as the only V&V method available for software, this is probably why you get confused. Testing is just one of many methods for V&V. Simple examples that come to my mind are mathematical proofs, model checking, static code analysis. All of this activities are part of the V&V process regardless which department in a company conducts them. But the passage is ok now, it's just that before it stated that testing is bigger than V&V, where in fact is a part of the V&V.--Stoilkov (talk) 10:34, 12 November 2011 (UTC)[reply]
I'm sorry. I don't think you understand. Just because it's part of one group doesn't mean it can't be part of another. The statement does not make it exclusive to software testing. If you would like to add other disciplines, feel free. However, static code analysis is part of the software testing process as well. The fact that it's done by developers doesn't mean that it's not a testing process. --Walter Görlitz (talk) 16:31, 12 November 2011 (UTC)[reply]
I give up :) This is the defintion of testing from the Software Engineering Body of Knowledge :

Software testing consists of the dynamic verification of the behavior of a program on a finite set of test cases, suitably selected from the usually infinite

executions domain, against the expected behavior.

If you thing this definition holds for static code analysis, model-checking and some other formal way of verification, this means you are probably right :)Stoilkov (talk) 11:39, 15 November 2011 (UTC)[reply]

Shall I offer my definitions for the half dozen books in my possession? Most, if not all, include static as well as dynamic review. --Walter Görlitz (talk) 15:02, 15 November 2011 (UTC)[reply]
I am sorry, I guess we are talking about different things here. I never said anything about reviews, my point was (and i gave you an example for this) that V&V is a process for which you can use different methods, including testing ... and reviews... and formal methods ... and others. Sure you can say "we test that a program is correct using the Hoare logic, so this is part of software testing". This means that you just use Testing as synonym of Verification. If this is your statement, i would surely would want to take a look at your sources. However your edit corrected the error, so i will close this discussion on my part Stoilkov (talk) 17:17, 15 November 2011 (UTC)[reply]

System testing

[edit]

IMHO, System testing is releated pages. — Preceding unsigned comment added by Michal Jurosz (talkcontribs) 06:46, 15 October 2013 (UTC)[reply]

Merge Proposal

[edit]

I propose turning the so-called "ISVV" article into a redirect to the relevant subsection of this other article. The other is both of low quality and redundant in light of this article. 67.161.69.249 (talk) 19:06, 26 September 2014 (UTC)[reply]

Support Walter Görlitz (talk) 23:53, 29 March 2017 (UTC)[reply]
 Done Klbrain (talk) 09:55, 8 July 2017 (UTC)[reply]

How do you call the person who does verification or validation?

[edit]

Is it a verficator and validator? Please add an answer to the article if possible. Thanks --Hfst (talk) 20:38, 23 July 2019 (UTC)[reply]

Move discussion in progress

[edit]

There is a move discussion in progress on Template talk:IEEE software documents which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 22:02, 15 June 2020 (UTC)[reply]

Move discussion in progress

[edit]

There is a move discussion in progress on Template talk:IEEE software documents which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 23:32, 15 June 2020 (UTC)[reply]