good response Paul. You made a lot of great points.

now get up off the floor and take a deep breath :)

Paul Replies...

Thanks.

Now challenge a proposed axiom or two!

:-)

As you already know I like the discussion started. In the last few days both postings, as well by you as by James made me think further. For me I came up with this conclusion: Axioms can be useful to explain to stakeholders/ management what we mean with testing. Axioms can help explaining to the team what we explained to management. Axioms should be used wise fully and with respect and in it own context.

And I also think your posting can help how to deal, use and respect them to make sure these Axioms are not used with wrong understanding.

with regards,
Jeroen

Paul Replies...

Thanks Jeroen. I'm not trying to put the world to rights. I certainly don't want to own the axioms - if they appear to be something useful - le'ts refine them as a community and use them to communicate what testers do as well as use them as thinking tools.

As I've said in another reply - maybe axioms IS the wrong term. But, by calling them axioms now, they pose a challenge to all of us - so I hope they can be refined And then maybe axioms WOULD be the right term. I really don't know how this will turnout eventually. It could be an interesting diversion - or something useful. Useful I hope!

Paul:
By the omission of the expected apostrophe in “School’s Out”, was that deliberate, are you subliminally requesting all schools to “out” (ie declare) themselves, or suggesting an ending of the “schools” distinction? There’s a blog entry by “I.M.Testy” (BJ Rollison), which seems to me to put that latter argument well:
http://blogs.msdn.com/imtesty/archive/2006/10/20/end-segregation.aspx

I’m not against the schools in principle, I think the distinction is useful. But please see also James Bach’s blog entry http://www.satisfice.com/blog/archives/74 in which James states that “attitudes [context-oblivious, context-specific, context-imperial & context-driven] are not mutually exclusive”; “these attitudes, unlike schools, can mix together”, and (in reply to a commenter) that “any school could use any practice”. So...
As you know, I have been (and still am) at least trying to persuade the schools to work more amenably together. This is so far turning out more difficult than I’d hoped!
And who, really, is in which school? Are we all wearing badges? I can imagine a client meeting which might go as follows:
• Client: we want you to test X in the context Y.
• I: Please tell me more about that context (or circumstances!), so I can best advise you.
• Client: blah blah.
• I: Ah, that sounds like you need practices 24, 193, 299... (for now, though it may change soon). That’s mostly available from school alpha-tau, but perhaps with some elements of school rho-tau!
• Client: Actually I’d heard that school beta-pi-tau is the best (by definition!), of course we want that.
• I: But the thing is, school chi-delta-tau asserts that there is no such thing, so you’ve got to choose between the others, or just let me choose all the practices for you individually. Or even better, use chi-delta-tau because it may be thought to subsume all the others.
• Client: Hmm, I’m not sure about that – and I’ve read James Bach’s blog and he has stated that it’s not like that, rather that “the schools are independent belief systems”. So... I really like the sound of that mixture of alpha-tau and rho-tau.
• I: This is difficult for me, because I’m not religious, I want everyone to learn from each other.
• Client: Do you want this job or not?
• I: Well yes, but I’m not accredited to belong to any of the individual schools, and I’m not mentally committed to any one of them, so I’ll have to refer you instead to specific people from among my extensive network of contacts. Will you make your *%@ing mind up which school you want?
• Client: No, *%@ off then.

So (readers all) please do not misunderstand me, I am not attacking any individual school, I am expressing sadness and fear at the apparent hostility between members of different “schools”.

Regards,
Neil.

Paul Replies...
Thanks Neil,

It's a typo. I checked on alice Cooper's website and it's spelt correctly now.

Rather than get into a long debate here, I've posted a blog on my view of testing schools here.

To respond directly to your comments above though...

The BJ Rollison post is excellent and hits it on the button.

I do notice though that Bret has added 'Standard' to his list of schools since that post.

I've never experienced anything that I would call hostility from anyone in the testing community - and I'm really not that bothered. They usually miss the point.

But thaving said that, I've never met anyone who said they were a member of ANY single school at all except the context-driven folk - but there's only a handful of those in the UK aren't there? The other schools have been defined by Bret. Is anyone else interested in schools apart from the CD folk?

I think people who align themselves with schools taint their marketability as consultants - but I can't worry about that.

.... end of Paul's comment...

>>> is anyone else interested in schools apart from the CD folk?

I think very few or nearly NONE ... do you know why ..

Like BJ.. many people who attack schools concept - at first declare that there are no schools. That is a sad, immediate end to an otherwise interesting story.

Declaring that there are no schools- you would be killing any further discusion. Also declaring that there are no schools - you ackowledge that there are no differences in testing practice.

Do you think that your system of axiom is a school in itself? or you also belive that schools do not exist (other than in the minds of CD thinkers)

Shrini

Paul ReplieS:

I made a robust rejection of the "Gerrard School" because it doesn't exist and I've no intention of starting it.

The notion of "all testing is context-driven" is pefectly sound. I've got no problem with that CD statement. (Although I could argue that all human thought and endeavour is context-driven - but that's the full 30 minute argument ;-)

But the CD school have defined other schools as you know. I'm not worried about whether the definition of these schools is exactly accurate or not - there's a good case for separating these out. But I do believe *all* testing is context-driven and that applies to safety-critical systems as much as shrink-wrapped software with a release deadline to meet. So this is a problem for me.

I'm not very worried about schools fragmenting our industry and causing conflict, antagonism etc. That's a people thing, not a school thing. I am worried that the schools' definitions doesn't seem to match my reality.

For example, I think there is a place for exploratory testing in every kind of software project. High integrity projects dominated by industrial strength contracts, huge expenditure, political and life-threatening risks, their need to be formal, transparent, systematic, planned to the nth degree etc. etc.

Stage payments may be based on the review of detailed test plans, execution and witnessing of testing and retesting, and review and sign-off of vast quantities of documentation. Now, is there value in setting aside some time for exploration? Of course there is. A bug found this way is as good as bugs found by any other means. But the crux of it to me is: exploratory testing is probably not contractual (other than you might say, e.g. spend XX mandays in YY elapsed weeks to do it).The test evidence provided by exploration, other than incident/defect reports, will not form part of the stage assessment for the approval of a stage payment because it can't really be planned for. No problem. Not contractual.

This is an extreme case, but I would suggest ET could be included in such a project with obvious benefits. As we move across the spectrum towards more let's say, Agile projects, it becomes easier and easier to fit in ET naturally. What about the industrial strength plan, spec, prepare, execute, analyse approach? Well it's clear that these approaches become less appropriate of course - so levels of formality naturally reduce. So it is a spectrum, I believe. Where does the range of context-driven testing span? Well it seems to be very focused on software projects where requirements are very poor, an agile-ish approach is already inplace etc. etc. That limits it's applicability which is a shame if it is intentional.

If I believe 'high process' approaches have their place in some projects - does that preclude me being context-driven? That is the very strong impression one gets.

I'm not hostile to schools of thought but the schools as defined by Bret don't help me - if a school is a school of thought - it is thought-limiting - BY DEFINITION. I know the CD guys promote better thinking processes and that is absolutely the way to go. But a continuing theme of these schools discussions is that high-formality approaches are the result of crooked thinking.

Projects and companies working at the high integrity end of the spectrum try and promote the notion that they are doing 'software engineering'. Although they might have industrial-strength (or industrial-weakness) processes, the guys doing testing are likely to be very smart people indeed. In the UK at least, there's been a tradition of software engineering companies hiring very smart people - maths, physics grads, (comp.science even) and phDs etc. Are these people flatfooted - no. Do these people use their brains and heuristics to construct tests - absolutely. Suggesting that the thinking that goes on in big projects is always flat-footed is just offensive. Unfortunately, there are an awful lot of undisciplined people who work in more laissez-faire environments because it attracts them.