--- In agile-testing@yahoogroups.com, Janet Gregory <janet_gregory@...> wrote:
>
> The 'whole team' approach is absolutely needed for
>successful agile projects. There is much more
>collaboration than "a hand-off" from developer to tester.
>In a truly collaborative environment, the developers have
>all the tests early to help them code.
>Once the code is complete, they will demo to the tester
>&/or customer. As Steve says, what is left at that time,
>is the exploratory testing that cannot necessarily be
>defined up front. Developers don't "hand off" since they
>are still involved until the story is marked done
>(testing is complete).
>

Janet, I like you, and, at this point, I feel I've been substantially
misrepresented.

I am not arguing against early QA involvement, nor am I suggesting that devs
start with a blank slate an no story tests up front.

I /am/ suggesting that many of the best test ideas occur after the code is
demoable, and that stories reaches a point where they need aggressive testing,
at, at the that point, it's best for the developer to involve a tester if the
team has them.

Stephen Gordon has suggested that is not a "whole team" approach, and, here, in
your words, your wrote:

"The 'whole team' approach is absolutely needed for successful agile projects"

Seeing as my team has shipped working software to production for 36 of the past
40 iterations, and "whole team" is "absolutely needed" for successful agile
projects, I declare my team is doing "whole team" and end the discussion.

regards,

--matthew heusser
(*) - The speaker could, of course, argue that Heusser's team is successful but
/not Agile/, the classic argument of the person who's run out of ideas.  Hey
man, we ship working software, we respond to change, we focus on individuals and
interactions.  Who's the one focusing on definitions and process? :-)

--- In agile-testing@yahoogroups.com, Steven Gordon <sgordonphd@...> wrote:
>
> Matthew,
>
> I do not understand why you go out of your way to misspell
>my name so often.
>

I can remember to make it Steve.  I honestly apologize.

Also, Steve, I accept your point that the article could be misunderstood.

It took me about 10,000 words to write my chapter for beautiful testing, because
anything less would miss something important.  Perhaps a 'scope statement' at
the top would help?

--heusser

> If you are delivering software successfully, I don't really care what process
you use... though I might care what you name  it :-)

Really? Why? If someone is delivering software successfully, I really do care
about the process they use, but couldn't care less about what they name it.

Andrew

--- In agile-testing@yahoogroups.com, Janet Gregory <janet_gregory@...> wrote:
>
> Matt,
>
> Point taken - I may have taken words out of context. I'll reread the article
again. There is definitely a need to do some kind of exploratory testing (as
well as other types of testing) after the code has been written, and and testers
are usually the best people to do it.
>
> If you are delivering software successfully, I don't really care what process
you use... though I might care what you name  it :-)
>
> Janet
>

I now use GoogleTest.  I used to use a simple homegrown framework that
I used when teaching C++.  The key factor for me with a test framework
is that the tests must be self registering.  Otherwise you risk
accidentally not running tests.

Other reasons that I moved to GoogleTest are:
- XML output in JUnit format that integrates nicely with Hudson for
continuous integration
- the ability to disable certain tests
- setup and teardown
- terminating and non-terminating assertions

Things I'd like to change or work out how to do with GT:
- get IDE-friendly error messages so that I can jump to the offending line

GoogleMock integrates well with GoogleTest.  Mocking in C++ requires
abstract base classes with pure virtual functions.  You may need to
insert some of these.  They are often a good thing but excessive
mocking is probably an indication of too fine-grained testing.  You
might also like to consider using templates instead of inheritance to
substitute types.  Function pointers are another trick along with
function objects (functors).

On Linux I have also used loader substitution of functions (the --wrap
option of ld) and LD_PRELOAD for mocking functions.

--
Hubert Matthews         http://www.oxyware.com/
Software Consultant     hubert@...

As an organizer of a small Agile user group and potentially a small Agile conference I’ve noticed that the Community lacks one good central calendar with a list of events. If we created one – would you find it useful?

I’ve suggested that InfoQ create one and need to determine what the minimum feature set would be:

• Events are tagged with their geographic location, time, date, cost – with everything else in free form text
• Only not for profit events are allowed I’m not quite sure what test to apply here I just know that we would drown in Scrum/Agile training announcements if they were permitted on the same calendar.
• Subscribe via RSS
• Subscribe via monthly/bi-weekly email.
• Moderated to prevent spam
• Anyone can submit an event

In the wine world something like this already exists: http://www.localwineevents.com/events/listing/all/76

My questions:

• Would this be useful to you?
• Would you subscribe?
• What features am I missing from the minimum feature set?
• Would you be willing to help moderate this?
• Is there any existing open source/free software that can be used to this job?
• Would you think to look on: http://planet.infoq.com/ to find the calendar?

		Mark Levison | Founder and Consultant - TheAgileConsortium | Agile Editor @ InfoQ Blog | Twitter | Office: (613) 761-9821 Recent Entries: Agile Mailing Lists a Survey, Do You Suspect You Have a Less than Productive Person on Your Team?

try:

AgileANN in Yahoo

Give THANKS for Scrum ! Jeff, Ken, Sanjiv, Amr, Dan
http://www.newtechusa.com/agileboston/gtfs



> As an organizer of a small Agile user
> group<http://agileottawa.wordpress.com/>and potentially a small Agile
> conference I’ve noticed that the Community
> lacks one good central calendar with a list of events. If we created one –
> would you find it useful?
>
> I’ve suggested that InfoQ create one and need to determine what the
> minimum
> feature set would be:
>
>    - Events are tagged with their geographic location, time, date, cost –
>    with everything else in free form text
>    - Only not for profit events are allowed *I’m not quite sure what test
> to
>    apply here I just know that we would drown in Scrum/Agile training
>    announcements if they were permitted on the same calendar.*
>    - Subscribe via RSS
>    - Subscribe via monthly/bi-weekly email.
>    - Moderated to prevent spam
>    - Anyone can submit an event
>
> In the wine world something like this already exists:
> http://www.localwineevents.com/events/listing/all/76
>
> My questions:
>
>    - Would this be useful to you?
>    - Would you subscribe?
>    - What features am I missing from the minimum feature set?
>    - Would you be willing to help moderate this?
>    - Is there any existing open source/free software that can be used to
>    this job?
>    - Would you think to look on: http://planet.infoq.com/ to find the
>    calendar?
>
> BTW I've seen Deb's efforts at: http://upcoming.yahoo.com/group/2228/  -
> but
> it doesn't appear to have taken off.
>
> Cheers
> Mark Levison
>
>    *Mark Levison* | Founder and Consultant -
> TheAgileConsortium<http://theagileconsortium.com/>| Agile
> Editor @ InfoQ <http://www.infoq.com/about.jsp>
> Blog <http://www.notesfromatooluser.com/> |
> Twitter<http://twitter.com/mlevison>| Office: (613) 761-9821
> Recent Entries: Agile Mailing Lists a
> Survey<http://www.notesfromatooluser.com/2009/06/agile-mailing-lists.html%20>,
> Do You Suspect You Have a Less than Productive Person on Your
>
Team?<http://www.notesfromatooluser.com/2009/01/do-you-suspect-you-have-a-less-t\
han-productive-person-on-your-team.html>
>

On Monday 26 October 2009 06:21:04 pm James Martin wrote:
> I recently switched from Boost Test to CppUTest and have been really
> impressed so far. The rationale was to break dependencies with Boost (which
> can be a little overkill at times). CppUTest is lightweight, does memory
> leak detection, compiles easily on nix/osx/windows and has some nice helper
> scripts/features bundled in for TDDers.
>
> http://cpputest.org.
>
> I'd advise grabbing SVN HEAD from sourceforge to try out the latest stuff,
> including ready-to-go example .project files for Eclipse.

I've spent most of the last few years working with Qt (qt.nokia.com), a cross
platform GUI toolkit.  Although I say GUI, it's modular can be used for non-GUI
applications as well.  It includes a unit test library called QTestLib that
works quite well (http://doc.trolltech.com/4.5/qtestlib-manual.html)

The main features that it includes that I like are:

* Basic GUI Testing (drive combo boxes, input fields, mouse clicks, etc.)
* Benchmarking
* Data-driven testing (run multiple data sets through a single test)

It does have some weaknesses in the test management area.  If you have hundreds
of tests, it's a pain to select which tests to run.  I wrote a wrapper around
QTestLib that allows me to specify the test class or classes, using a regex if
desired, and then the tests to run.  Lastly, it's xml output option was useful
for integrating with a CI-server so I could get feedback about the unit tests.

I've used QTestlib for testing C projects that I've worked on as well.

I've been tempted to play with googletest
(http://code.google.com/p/googletest/). Does anybody have any experience with
it?

Thanks.

--
Kaleb Pederson

Twitter - http://twitter.com/kalebpederson
Blog - http://kalebpederson.com

On Tuesday 27 October 2009 08:54:17 am Mark Levison wrote:
> My questions:
>
>    - Would this be useful to you?

Yes.

>    - Would you subscribe?

Yes if ... (see below)

>    - What features am I missing from the minimum feature set?

I would want to narrow down events to a specific geographical region.

>    - Would you be willing to help moderate this?

Yes.

>    - Is there any existing open source/free software that can be used to
>    this job?
>    - Would you think to look on: http://planet.infoq.com/ to find the
>    calendar?

No, but as long as my it showed up in my RSS reader I'd be able to get to it.

--
Kaleb Pederson

Twitter - http://twitter.com/kalebpederson
Blog - http://kalebpederson.com

Hi,

Hemal Kuntawala delivered a fantastic experience report yesterday at the
agile testing user group meeting, talking about all the process
improvements implemented in his company over the last year. The changes
were driven by a need for better quality but ended up fundamentally
changing the way they think about software and deliver it. Here's a
write up, it should prove to be a very interesting read if you weren't
there to see the talk in person.

http://gojko.net/2009/10/29/upgrading-agile-development-at-uswitch-com-from-conc\
ept-to-production-in-four-days/

--
gojko adzic
http://gojko.net

Argh! So much to read, so few time.... I put it in my reading queue....

Kind regards
Markus Gärtner

Lisa Crispin wrote:
> I loved reading about this company's efforts and how it paid off, thank you
> for writing it up!
> -- Lisa
>
> On Thu, Oct 29, 2009 at 1:47 PM, Gojko Adzic <gojko-yahoolist@...>wrote:
>
>>
>> Hi,
>>
>> Hemal Kuntawala delivered a fantastic experience report yesterday at the
>> agile testing user group meeting, talking about all the process
>> improvements implemented in his company over the last year. The changes
>> were driven by a need for better quality but ended up fundamentally
>> changing the way they think about software and deliver it. Here's a
>> write up, it should prove to be a very interesting read if you weren't
>> there to see the talk in person.
>>
>>
>>
http://gojko.net/2009/10/29/upgrading-agile-development-at-uswitch-com-from-conc\
ept-to-production-in-four-days/
>>
>> --
>> gojko adzic
>> http://gojko.net
>>
>>
>
>
>

I'm too lazy to look up the references, but there's a generally accepted
position that any reasonable agile process will be at least CMM Level 3
compliant.

There is a less generally-accepted position that pursuing CMM Level 4 and 5
don't provide much benefit except in certain very rare cases.

I'm sure google will help you out given these pointers.

I see less and less attention paid to CMM/Six Sigma/ISO9000 from the agile
community.  Slavishly following any of them simply becomes damage to be routed
around.  See the Fishing Maturity Model.



--- In agile-testing@yahoogroups.com, Heba Hosny <heba.hosny@...> wrote:
>
> Hi all,
>
> I was wondering if Agile &  CMMi can be merged in some way, I'm going to
> attend a conference that highlight this but first I'd like to know more of
> your viewpoints.
>
> I found that <http://www.sei.cmu.edu/library/abstracts/reports/08tn003.cfm>
> about
> this topic.
>
> Waiting for your replies,
>
> --
> All the best,
> Heba Hosny
> -----------------------
> Never test the depth of a river with both feet
>

--- In agile-testing@yahoogroups.com, Lisa Crispin <lisa.crispin@...> wrote:
>
>
> There's nothing incompatible between agile and
> any of the quality models.  Coordinate with the
> folks in your company that deal with the quality model,
> and you should be able to work it all out. I've talked
> to more ISO shops that do agile than CMMI, but it's
> all the same kinda thing.
> -- Lisa

At the lower levels, CMMI says "document everything."  At the higher levels, it
says "measure everything."

There are significant issues in cognitive science and knowledge work with trying
to document everything(1).  There are significant behavioral, social, and
construct validity issues with trying to measure everything(2,3).

The CMMI also fails to deal with key dynamics of software development - most
importantly, the value of rapid feedback.  While reviewing every "work product"
is constantly mentioned, the idea to get to working software quickly really
isn't a core value of the model.

The folks at the SEI respond to these problems, for the most part, by ignoring
them.  Search the SEI website for "dysfunction" and you'll get all of two hits -
critical issues in software measurement.  They are not responded to.

Moreover, the CMMI documents themselves were poorly written. You can tell this
OBJECTIVELY, with metrics, by evaluating the Flesch-Kincaid readability score of
the documents(4).

They generally score so high as to be incomprehensible.  This is not my
/opinion/; it is repeatable experiment.  I've got /metrics/.

What you left with, after all that, is the fishing maturity model:

http://blogs.stpcollaborative.com/matt/2009/10/08/the-fishing-maturity-model/

Bottom Line: If you must do CMMI, I would consider it a wasteful-but-required
process, in the Lean sense.


Regards,

--heusser
blog: http://blogs.stpcollaborative.com/matt
References:

(1) - "The Social Life of Information" -
http://www.amazon.com/gp/product/1578517087?ie=UTF8&tag=heusseronlead-20&linkCod\
e=as2&camp=1789&creative=390957&creativeASIN=1578517087

(2) - "Software Engineering Metrics: What do they measure and how do we know?"
Kaner and Bond - http://www.kaner.com/pdfs/metrics2004.pdf

(3) - "Measuring and Managing Performance in Organizations", Robert Austin -
http://www.amazon.com/gp/product/0932633366?ie=UTF8&tag=heusseronlead-20&linkCod\
e=as2&camp=1789&creative=390957&creativeASIN=0932633366

(4) - http://en.wikipedia.org/wiki/Flesch%E2%80%93Kincaid_readability_test

I've corrected my opinion of the subject line.

--heusser

Hi,

A great write-up, Gojko.. Many thanks! :)

I've written a quick note linking to the slides and resources on my blog:

http://testerhemal.wordpress.com/2009/10/30/slides-from-how-we-build-quality-sof\
tware-at-uswitch-com/

Thanks again for having me, Gojko. I look forward to next month's event. (Thanks
for your help and support too, Aidy :) ).



Cheers,
Hemal



--- In agile-testing@yahoogroups.com, aidy lewis <aidy.lewis@...> wrote:
>
> It is fully gratifying that uSwitch accepted my ideas on early Cucumber and
> Watir a number of months ago,   when others were writing testing books
> promoting system languages and Selenium,
>
> Aidy Lewis
>
>
> 2009/10/29 Gojko Adzic <gojko-yahoolist@...>
>
> >
> >
> > Hi,
> >
> > Hemal Kuntawala delivered a fantastic experience report yesterday at the
> > agile testing user group meeting, talking about all the process
> > improvements implemented in his company over the last year. The changes
> > were driven by a need for better quality but ended up fundamentally
> > changing the way they think about software and deliver it. Here's a
> > write up, it should prove to be a very interesting read if you weren't
> > there to see the talk in person.
> >
> >
> >
http://gojko.net/2009/10/29/upgrading-agile-development-at-uswitch-com-from-conc\
ept-to-production-in-four-days/
> >
> > --
> > gojko adzic
> > http://gojko.net
> >
> >
>

--- In agile-testing@yahoogroups.com, "Hagar, Jon D" <jon.d.hagar@...> wrote:
>
> ... Lisa is right.  I have worked at places that were
> both some flavor of Agile and CMMI level 3 (independently
> assessed).
>

The CMMI is a bunch of words.  Of course you can 'do' it, for some definition of
'do'. You might have to in order to win a contract.  I'm saying as a process
improvement framework, it's an oxymoron.

I grant that you /can/ get /some/ value out of /some/ aspects of the CMMI.  I
would add that if you're advanced enough to be able to pull this value out, you
don't need the models: Just read Jerry Weinberg's Quality Software Management:
Volume I and get to it.


--heusser