Why is UI testing at the top of the test pyramid? - selenium

I've been asked by an interviewer the following questions about the diagram below:
Why should less effort be required for UI testing?
Is the test pyramid meant for programmers or testers?
Thanks for your help on these questions.

My answer for 1. is that it does not mean "less effort" should be put into UI testing, but that "less code should be covered" by UI testing than the larger portions.
As for 2., it should by considered by anybody who is testing the code. In some organizations that is (and should be) most often a specialized tester and in others, a developer.
My impression is that these are not very good questions, but I do not mention this during the interview.

As per Test Pyramid: the key to good automated test strategy the following diagram represents the Test Pyramid which was originally reflected in one of Martin Fowler's article which was based on a concept developed by Mike Coen.
Test Pyramid
The test pyramid is a tool to fix the problem of over-reliance on long-running Automated / Manual UI tests.
Take aways from the TestPyramid
The pyramid implies that:
Tests on the lower levels are cheaper to write and maintain, and quicker to run.
Tests on the upper levels are more expensive to write and maintain, and slower to run.
Ideally, you should have lots of unit tests, some service tests, and very few UI tests.
This concept is exactly the reverse case of the Testing Ice Cream Cone which is the most often found Test Strategy followed across the software industry where we have very few unit tests, some UI test, lots of QA tests and lots of Manual Tests.
Testing Ice Cream Cone
The Testing Ice Cream Cone approach can be easily identified where the QA department has created an automated test suite, but the development team has not. As a result, tests will be very long running and flakey because the development team has not helped build the suite or architect the application in a way that makes it easy to test. The new check-ins by the developers breaks the product functionality regularly and they are relying on the QA department to report the defects to fix it up.

Related

Is Functional Automation Testing worth the effort?

Do you think that the effort which we put to automated functional tests really impacts?
As most of the time it considered as secondary option in many service based companies and when it comes to delivery, they rely on Manual testing.
Short answer
Yes
Long answer (short version)
Depends
if you have budget and time
if you have people with necessary skills
if you maintain the project in support or if this is a long project in order to reuse the automation suite
how big the project is - in case you don't have enough time later to rerun all manual tests between important releases because there are too many features to cover or you don't have enough people
you have repetitive tasks, project is stable and you don't have major changes with each release
I would say: Automate if you feel it brings value, this is a decision based on each project and team.
You have a lot of tools to choose from, even easy ones with record and play if you need a small suite and you don't have the skills for something complex.
Automated Functional Test Really Impacts?
I would say Yes.
However, its coverage is not as wide as manual testing. Service based company will prefer to use manual testing as they will need to perform a lot of exploratory testing. Just to make sure that there's no introduced issue on the other (possibly) affected area. In my opinion, that kind of testing is not fully supported by automated testing. Unless you have a lot of test cases - which might affect the software maintenance/development cost.
I think regression testing is good for regression testing - to avoid repetitive works for software tester. As in my experience, regression testing will executing the same end-to-end test case.

Can someone explain differences of Test-Driven Development, Agile Development, SCRUM & Unit Testing

And further, how they relate together or even if they do?
What would one do to understand the various pieces to a simple question, how to I properly build a testing facility for my (web or other) application?
Agile Development is a banner term for many things, too numerous to mention, including Scrum and TDD. It typically, but not always follows the Agile Manifesto.
SCRUM
This is a particular flavour of agile. This diagram from wikipedia shows the process:
See wikipedia for more info.
Unit Testing
This is the art of writing code that tests code. Failing tests indicate a problem in your solution.
Test Driven Development
This is the practice of writing tests before code, some of the advantages being that untested code isn't added to the solution, and that the code written is testable.
A proper testing facility, usually leverages something along the lines of xunit, junit, nunit, mstest depending on the framework used, these tests are typically ran via a Continuous Integration build on some kind of build server. That is a build that runs everytime the code changes, that executes tests. This way problems are identified quicker.

Difference between acceptance test and functional test?

What is the real difference between acceptance tests and functional tests?
What are the highlights or aims of each? Everywhere I read they are ambiguously similar.
In my world, we use the terms as follows:
functional testing: This is a verification activity; did we build a correctly working product? Does the software meet the business requirements?
For this type of testing we have test cases that cover all the possible scenarios we can think of, even if that scenario is unlikely to exist "in the real world". When doing this type of testing, we aim for maximum code coverage. We use any test environment we can grab at the time, it doesn't have to be "production" caliber, so long as it's usable.
acceptance testing: This is a validation activity; did we build the right thing? Is this what the customer really needs?
This is usually done in cooperation with the customer, or by an internal customer proxy (product owner). For this type of testing we use test cases that cover the typical scenarios under which we expect the software to be used. This test must be conducted in a "production-like" environment, on hardware that is the same as, or close to, what a customer will use. This is when we test our "ilities":
Reliability, Availability: Validated via a stress test.
Scalability: Validated via a load test.
Usability: Validated via an inspection and demonstration to the customer. Is the UI configured to their liking? Did we put the customer branding in all the right places? Do we have all the fields/screens they asked for?
Security (aka, Securability, just to fit in): Validated via demonstration. Sometimes a customer will hire an outside firm to do a security audit and/or intrusion testing.
Maintainability: Validated via demonstration of how we will deliver software updates/patches.
Configurability: Validated via demonstration of how the customer can modify the system to suit their needs.
This is by no means standard, and I don't think there is a "standard" definition, as the conflicting answers here demonstrate. The most important thing for your organization is that you define these terms precisely, and stick to them.
I like the answer of Patrick Cuff. What I like to add is the distinction between a test level and a test type which was for me an eye opener.
test levels
Test level is easy to explain using V-model, an example:
Each test level has its corresponding development level. It has a typical time characteristic, they're executed at certain phase in the development life cycle.
component/unit testing => verifying detailed design
component/unit integration testing => verifying global design
system testing => verifying system requirements
system integration testing => verifying system requirements
acceptance testing => validating user requirements
test types
A test type is a characteristics, it focuses on a specific test objective. Test types emphasize your quality aspects, also known as technical or non-functional aspects. Test types can be executed at any test level. I like to use as test types the quality characteristics mentioned in ISO/IEC 25010:2011.
functional testing
reliability testing
performance testing
operability testing
security testing
compatibility testing
maintainability testing
transferability testing
To make it complete. There's also something called regression testing. This an extra classification next to test level and test type. A regression test is a test you want to repeat because it touches something critical in your product. It's in fact a subset of tests you defined for each test level. If a there's a small bug fix in your product, one doesn't always have the time to repeat all tests. Regression testing is an answer to that.
The difference is between testing the problem and the solution. Software is a solution to a problem, both can be tested.
The functional test confirms the software performs a function within the boundaries of how you've solved the problem. This is an integral part of developing software, comparable to the testing that is done on mass produced product before it leaves the factory. A functional test verifies that the product actually works as you (the developer) think it does.
Acceptance tests verify the product actually solves the problem it was made to solve. This can best be done by the user (customer), for instance performing his/her tasks that the software assists with. If the software passes this real world test, it's accepted to replace the previous solution. This acceptance test can sometimes only be done properly in production, especially if you have anonymous customers (e.g. a website). Thus a new feature will only be accepted after days or weeks of use.
Functional testing - test the product, verifying that it has the qualities you've designed or build (functions, speed, errors, consistency, etc.)
Acceptance testing - test the product in its context, this requires (simulation of) human interaction, test it has the desired effect on the original problem(s).
The answer is opinion. I worked in a lot of projects and being testmanager and issuemanager and all different roles and the descriptions in various books differ so here is my variation:
functional-testing: take the business requirements and test all of it good and thorougly from a functional viewpoint.
acceptance-testing: the "paying" customer does the testing he likes to do so that he can accept the product delivered. It depends on the customer but usually the tests are not as thorough as the functional-testing especially if it is an in-house project because the stakeholders review and trust the test results done in earlier test phases.
As I said this is my viewpoint and experience. The functional-testing is systematic and the acceptance-testing is rather the business department testing the thing.
Audience. Functional testing is to assure members of the team producing the software that it does what they expect. Acceptance testing is to assure the consumer that it meets their needs.
Scope. Functional testing only tests the functionality of one component at a time. Acceptance testing covers any aspect of the product that matters to the consumer enough to test before accepting the software (i.e., anything worth the time or money it will take to test it to determine its acceptability).
Software can pass functional testing, integration testing, and system testing; only to fail acceptance tests when the customer discovers that the features just don't meet their needs. This would usually imply that someone screwed up on the spec. Software could also fail some functional tests, but pass acceptance testing because the customer is willing to deal with some functional bugs as long as the software does the core things they need acceptably well (beta software will often be accepted by a subset of users before it is completely functional).
Functional Testing: Application of test data derived from the specified functional
requirements without regard to the final program structure. Also known as
black-box testing.
Acceptance Testing: Formal testing conducted to determine whether or not a system satisfies its acceptance criteria—enables an end user to determine whether or not to
accept the system.
In my view the main difference is who says if the tests succeed or fail.
A functional test tests that the system meets predefined requirements. It is carried out and checked by the people responsible for developing the system.
An acceptance test is signed off by the users. Ideally the users will say what they want to test but in practice it is likely to be a sunset of a functional test as users don't invest enough time. Note that this view is from the business users I deal with other sets of users e.g. aviation and other safety critical might well not have this difference,
Acceptance testing:
... is black-box testing performed on a system (e.g. software, lots of manufactured mechanical parts, or batches of chemical products) prior to its delivery.
Though this goes on to say:
It is also known as functional testing, black-box testing, release acceptance, QA testing, application testing, confidence testing, final testing, validation testing, or factory acceptance testing
with a "citation needed" mark.
Functional testing (which actually redirects to System Testing):
conducted on a complete, integrated system to evaluate the system's compliance with its specified requirements. System testing falls within the scope of black box testing, and as such, should require no knowledge of the inner design of the code or logic.
So from this definition they are pretty much the same thing.
In my experience acceptance test are usually a subset of the functional tests and are used in the formal sign off process by the customer while functional/system tests will be those run by the developer/QA department.
Acceptance testing is just testing carried out by the client, and includes other kinds of testing:
Functional testing: "this button doesn't work"
Non-functional testing: "this page works but is too slow"
For functional testing vs non-functional testing (their subtypes) - see my answer to this SO question.
The relationship between the two:
Acceptance test usually includes functional testing, but it may include additional tests. For example checking the labeling/documentation requirements.
Functional testing is when the product under test is placed into a test environment which can produce variety of stimulation (within the scope of the test) what the target environment typically produces or even beyond, while examining the response of the device under test.
For a physical product (not software) there are two major kind of Acceptance tests: design tests and manufacturing tests. Design tests typically use large number of product samples, which have passed manufacturing test. Different consumers may test the design different ways.
Acceptance tests are referred as verification when design is tested against product specification, and acceptance tests are referred as validation, when the product is placed in the consumer's real environment.
They are the same thing.
Acceptance testing is performed on the completed system in as identical as possible to the real production/deployement environment before the system is deployed or delivered.
You can do acceptance testing in an automated manner, or manually.

Is there a right way to implement a continuous improvement (AKA software hardening) process?

Each release it seems that our customers find a few old issues with our software. It makes it look like every release has multiple bugs, when in reality our new code is generally solid.
We have tried to implement some additional testing where we have testers do several hours of monthly regression testing on a single app each month in an effort to stay ahead of small issues. We refer to this process as our Software Hardening process, but it does not seem like we are catching enough of the bugs and it feels like a very backburner process since there is always new code to write.
Is there a trick to this kind of testing? Do I need to target one specific feature at a time?
When you develop your testing procedures, you may want to implement these kind of tests:
unit testing (testing invididual components of your project to test their functionality), these tests are important because they allow you to pinpoint where in the software the error may come from. Basically in these tests you will test a single functionality and use mock objects to simulate the behavior, return value of other objects/entities.
regression testing, which you mentioned
characterization testing, one example could be running automatically the program on automatically generated input (simulating the user input), storing the results and compare the results of every version against these results.
At the beginning this will be very heavy to put in place, but with more releases and more bugs fixes being added to the automated non-regression tests, you should be starting to save time.
It is very important that you do not fall in the trap of designing huge numbers of dumb tests. Testing should make your life easier, if you spend too much time understanding where the tests have broken you should redesign the tests such as they give you better messages/understanding of the problem so you can locate the issue quickly.
Depending of your environment, these tests can be linked to the development process.
In my environment, we use SVN for versioning, a bot runs the tests against every revision and returns the failing tests and messages with the name of the revision which broke it and the contributor (his login).
EDIT:
In my environment, we use a combination of C++ and C# to deliver analytics used in Finance, the code was C++ and is quite old while we are trying to migrate the interfaces toward C# and keep the core of the analytics in C++ (mainly because of speed requirements)
Most of the C++ tests are hand-written unit tests and regression tests.
On the C# side we are using NUnit for unit testing. We have a couple of general tests.
We have a 0 warnings policy, we explicitely forbid people to commit code which is generating warnings unless they can justify why it is useful to bypass the warning for this part of the code. We have as well conventions about exception safety, the use of design patterns and many other aspects.
Setting explicitely conventions and best practices is another way to improve the quality of your code.
Is there a trick to this kind of testing?
You said, "we have testers do several hours of monthly regression testing on a single app each month in an effort to stay ahead of small issues."
I guess that by "regression testing" you mean "manually exercising old features".
You ought to decide whether you're looking for old bugs which have never been found before (which means, running tests which you've never run before); or, whether you're repeating previously-run tests to verify that previously-tested functionality is unchanged. These are two opposite things.
"Regression testing" implies to me that you're doing the latter.
If the problem is that "customers find a few old issues with our software" then either your customers are running tests which you've never run before (in which case, to find these problems you need to run new tests of old software), or they're finding bugs which you have previous tested and found, but which you apparently never fixed after you found them.
Do I need to target one specific feature at a time?
What are you trying to do, exactly:
Find bugs before customers find them?
Convince customers that there's little wrong with the new development?
Spend as little time as possible on testing?
Very general advice is that bugs live in families: so when you find a bug, look for its parents and siblings and cousins, for example:
You might have this exact same bug in other modules
This module might be buggier than other modules (written by somone on an off day, perhaps), so look for every other kind of bug in this module
Perhaps this is one of a class of problems (performance problems, or low-memory problems) which suggests a whole area (or whole type of requirement) which needs better test coverage
Other advice is that it's to do with managing customer expectations: you said, "It makes it look like every release has multiple bugs, when in reality our new code is generally solid" as if the real problem is the mistaken perception that the bug is newly-written.
it feels like a very backburner process since there is always new code to write
Software develoment doesn't happen in the background, on a burner: either someone is working on it, or they're not. Management must to decide whether to assign anyone to this task (i.e. look for existing previously-unfound bugs, or fix-previously-found-but-not-yet-reported bugs), or whether they prefer to concentrate on new development and let the customers do the bug-detecting.
Edit: It's worth mentioning that testing isn't the only way to find bugs. There's also:
Informal design reviews (35%)
Formal design inspections (55%)
Informal code reviews (25%)
Formal code inspections (60%)
Personal desk checking of code (40%)
Unit test (30%)
Component test (30%)
Integration test (35%)
Regression test (25%)
System test (40%)
Low volume beta test (<10 sites) (35%)
High-volume beta test (>1000 sites) (70%)
The percentage which I put next to each is a measure of the defect-removal rate for each technique (taken from page 243 of McConnel's Software Estimation book). The two most effective techniques seem to be formal code inspection, and high-volume beta tests.
So it might be a good idea to introduce formal code reviews: that might be better at detecting defects than black-box testing is.
As soon as your coding ends, first you should go for the unit testing. THere you will get some bugs which should be fixed and you should perform another round of unit testing to find if new bugs came or not. After you finish Unit testing you should go for functional testing.
YOu mentioned here that your tester are performing regression testing on a monthly basis and still there are old bugs coming out. So it is better to sit with the tester and review the test cases as i feel that they need to be updated regularly. Also during review put stress on which module or functionality the bugs are coming. Stress on those areas and add more test cases in those areas and add those in your rgression test cases so that once new build comes those test cases should be run.
YOu can try one more thing if your project is a long term one then you should talk with the tester to automate the regression test cases. It will help you to run the test cases at off time like night and in the next day you will get the results. Also the regression test cases should be updated as the major problem comes when regression test cases are not updated regularly and by running old regression test cases and new progression test cases you are missing few modules that are not tested.
There is a lot of talk here about unit testing and I couldn't agree more. I hope that Josh understands that unit testing is a mechanized process. I disagree with PJ in that unit tests should be written before coding the app and not after. This is called TDD or Test Driven Development.
Some people write unit tests that exercise the middle tier code but neglect testing the GUI code. That is imprudent. You should write unit tests for all tiers in your application.
Since unit tests are also code, there is the question of QA for your test suite. Is the code coverage good? Are there false positives/negatives errors in the unit tests? Are you testing for the right things? How do you assure the quality of your quality assurance process? Basically, the answer to that comes down to peer review and cultural values. Everyone on the team has to be committed to good testing hygiene.
The earlier a bug is introduced into your system, the longer it stays in the system, the harder and more costly it is to remove it. That is why you should look into what is known as continuous integration. When set up correctly, continuous integration means that the project gets compiled and run with the full suite of unit tests shortly after you check in your changes for the day.
If the build or unit tests fail, then the offending coder and the build master gets a notification. They work with the team lead to determine what the most appropriate course correction should be. Sometimes it is just as simple as fix the problem and check the fix in. A build master and team lead needs to get involved to identify any overarching patterns that require additional intervention. For example, a family crisis can cause a developer's coding quality to bottom out. Without CI and some managerial oversight, it might take six months of bugs before you realize what is going on and take corrective action.
You didn't mention what your development environment is. If yours were a J2EE shop, then I would suggest that you look into the following.
CruiseControl for continuous integration
Subversion for the source code versioning control because it integrates well with CruiseControl
Spring because DI makes it easier to mechanize the unit testing for continuous integration purposes
JUnit for unit testing the middle tier
HttpUnit for unit testing the GUI
Apache JMeter for stress testing
Going back and implementing a testing strategy for (all) existing stuff is a pain. It's long, it's difficult, and no one will want to do it. However, I strongly recommend that as a new bug comes in, a test be developed around that bug. If you don't get a bug report on it, then either is (a) works or (b) the user doesn't care that it doesn't work. Either way, a test is a waste of your time.
As soon as its identified, write a test that goes red. Right now. Then fix the bug. Confirm that it is fixed. Confirm that the test is now green. Repeat as new bugs come in.
Sorry to say that but maybe you're just not testing enough, or too late, or both.

Role of Testers in Agile? [closed]

Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 5 years ago.
Improve this question
I work in a team which has been doing the traditional waterfall method of development for many years. Recently, we've been told that future projects are going to be moving towards an agile (particularly Scrum) methodology. It so happens that my project will be one of the first, so we will essentially be guinea pigs for the next few months to iron out what it takes to make the transition.
The project itself is in a very early stage and we would usually be many months away from releasing anything to the testing team, but now we are going to be working directly with them up front. As a result, I'm concerned as to the role of the testers in such a project at this stage. I have several questions/concerns which hopefully some experienced agile developers could answer:
While a developer is coding a task, it is impossible for a tester to test it (it doesn't exist yet). What then is the role of a tester at this point
Is the tester now involved in unit testing? Is this done parallel to black box testing?
What does the tester do during a sprint where primarily infrastructural changes have been made, that may only be testable in unit testing?
How do the traditional test team members function in your agile project?
Keeping testers busy tends to get easier as a project matures (there is more to test!), but the following points apply in the early stages too:
Testers can prepare their test plans, test cases, and automated tests for the user stories before (or while) they are implemented. This helps the team discover any inconsistency or ambiguity in the user stories even before the developers write any code.
In my personal experience, testers don't have any involvement in unit testing; they only test code that passes all of the automated unit, integration and acceptance tests, which are all written by the developers. This split may be different elsewhere, though; for example your testers could be writing automated acceptance tests. Unit tests really should be written by the developers, however, as they are written in tandem with the code.
Their workload will vary between sprints, but regression tests still need to be run on these changes...
You may also find that having the testers spend the first couple of days of each sprint testing the tasks from the previous sprint may help, however I think it's better to get them to nail down the things that the developers are going to be working on by writing their test plans.
Ideally QA and testers should be involved if not from the day one then from very early stages of a software development project, regardless of the process used (waterfall or agile). The test team will need to:
Ensure that project or sprint requirements are clear, measurable and testable. In an ideal world each requirement will have a fit criterion written down at this stage. Determine what information needs to be automatically logged to troubleshoot any defects.
Prepare a project specific test strategy and determine which QA steps are going to be required and at which project stages: integration, stress, compatibility, penetration, conformance, usability, performance, beta testing etc. Determine acceptable defect thresholds and work out classification system for defect severity, specify guidelines for defect reporting.
Specify, arrange and prepare test environment: test infrastructure and mock services as necessary; obtain, sanitise and prepare test data; write scripts to quickly refresh test environment when necessary; establish processes for defect tracking, communication and resolution; prepare for recruitment or recruit users for beta, usability or acceptance testing.
Supply all the relevant information to form project schedule, work break down structure and resource plan.
Write test scripts.
Bring themselves up to speed with the problem domain, system AS-IS and proposed solution.
Usually this is not a question of whether a test team may provide any useful input into the project on an early stage, nor if such an input is beneficial. It is a question, however, of the extent to which an organisation can afford the aforementioned activities. There is always a trade off between available time, budget and resource versus the level of known quality of the end result.
Good post. I was in the same situation about 3 years ago and the transition from waterfall to agile was tricky. I encountered many pain points in the move but once I overcame them and my role had changed I realised that this way of working really suits testing.
The common myth that testers are not required is easily dispelled.
1. While a developer is coding a task, it is impossible for a tester to test it (it doesn't exist yet). What then is the role of a tester at this point
In my experience the tester could be working with the customer to fine tune the stories in the sprint.
They are usually working with the developers to fine tune the code that they are delivering. i.e. advising on edge cases, flows, errors etc.
They can often be involved in designing the tests that the coder will write to perform TDD.
If the agile team is fairly advanced then the tester would normally be writing the ATDD (Acceptance Test Driven Development) tests. These could be in a tool such as Fitnesse or Robot Framework or they could be more advanced ruby tests or even some other programming language. Or in some cases, simple record and playback can often be beneficial for a small number of tests.
They would obviously be writing tests and planning some exploratory testing scenarios or ideas.
The tricky thing to comprehend sometimes for the team is that the story does not have to be complete in order to drop it to the test stack for testing. For example the coders could drop a screen with half of the fields planned on it. The tester could test this half whilst the other half is being coded and hence feedback in with early test results. Testing doesn't have to take place on "finished" stories.
2. Is the tester now involved in unit testing? Is this done parallel to black box testing?
Ideally the coders would be doing TDD. Writing the test and then writing the code to make the test pass. And if the coders are wanting really good TDD then they would be liasing with the tester to think up the tests.
If TDD is not being done then the coders should be writing unit tests at the same time as coding. It probably shouldn't be an after thought or after task after the software has been dropped. The whole point of tests is to test the software is correct to avoid wasting time later down the line. It's all about instant feedback.
3. What does the tester do during a sprint where primarily infrastructural changes have been made, that may only be testable in unit testing?
Ideally the tester would be working with the team and the customer (who by the way, is part of the team!) to define the planned stories and build in some good, detailed acceptance critiera. This is invaluable and can save loads of time later down the line. The tester could also be learning new automation techniques, planning test environments, helping to document the outcome of the planning.
Ideally each story in the sprint would be testable in some way, shape or form. This doesn't mean it should be by the test team, but should be testable. So the tester could be working with the rest of the team working out how to make sure stories are testable.
I post some agile tips here : http://thesocialtester.posterous.com/
Hope this helps you out
Rob..
Just a few thoughts, definitely incomplete:
While the developer is coding a task, the tester can be examining the specifications (or requests from the customer, if there are no formal specs) and writing the test plan. This can include a conceptual framework for what needs to be tested, but it should also include formally writing test suites (yes, in code) as well. This can be quite a challenge for teams moving to agile, as a lot of testers are hired without programming skills. (In a lot of places, it seems like it's a requirement to not be able to code.)
The tester can be involved in unit testing, or in a slightly higher scope by testing components or libraries that have a clean interface.
The testers should always be executing regression tests, load tests, and any other kinds of tests that he can think of, as well as writing test suites for the next sprint. It's often the case that testers work one sprint ahead of development (in preparing a test environment), as well as one sprint behind development (in testing what developers just produced).
I saw a good talk on this recently. Basically this team started off doing a fairly standard Scrum process, then transitioned to Kanban and Lean. One of the most important things they did was to gradually erode the distinctions between testers and developers. Testers were involved in writing unit tests and code, developers were bringing in more higher level tests early in development. It was a steep learning curve for the testers, but worth it as the team was building in quality from the start. By now the testers call themselves developers because their work is so integrated in the process of writing code.
At my company we use and endorse Agile. Our QA team members are involved in unit test creation, maintaining the regression testing infrastructure and, just like in waterfall, they also test each feature upon completion.
When doing infrastructural changes, they also participate to make sure that the new infrastructure is testable.
So, from my limited experience, I'll try to answer your points:
If there's nothing to test yet, start setting up a regression/testing infrastructure and make sure that whatever is being done will be testable
Yes, he may do both
Maintains the testing infrastructure and hunts whoever breaks the tests
The most natural approach to testing in an agile environment is in my opinion exploratory testing http://en.wikipedia.org/wiki/Exploratory_testing.
Doesn't sound words like
According to Cem Kaner & James Bach, exploratory testing is more a [mindset] or "...a way of thinking about testing" than a methodology
or
pair testing
sound familiar to agile developpers. Testers can be involved much earlier in the process than in traditional testing.
1) While a developer is coding a task, it is impossible for a tester to test
it (it doesn't exist yet). What then
is the role of a tester at this point
The tester may still create test plans and have a list of what tests will be created. There may also be the need for the tester to get training if the development involves some off-the-shelf software,e.g. if you are doing a CMS project with Sitecore then the tester should know a few things about Sitecore. There can also be some collaboration of the tester, the developer and the end user or BA to know what are the requirements and expectations so that there isn't the finger pointing that can pop up in vague requirements.
2) Is the tester now involved in unit testing? Is this done parallel to
black box testing?
Not in our case. The tester is doing more integration/user acceptance testing rather than the low-level unit testing. In our case, unit tests come before any QA tests as the developers creating the functionality will create a layer of tests.
3) What does the tester do during a sprint where primarily infrastructural
changes have been made, that may only
be testable in unit testing?
Regression testing! In making infrastructural changes, did anything break? How thorough a testing suite can developers run compared to QA? We had this in a sprint not that long ago where most of the sprint work was plumbing rework so there wasn't much to test other than seeing that things that worked before still work afterward.
In our case, we have testing as one level up from our development environment but still a pre-production environment. The idea is to allow QA a sprint to validate the work done and for any critical or high severity bugs to be found and fixed before a release into staging for final user acceptance testing, so if developers are working on sprint X then QA is validating sprint X-1 and production may have sprint X-2 or earlier running depending on the final UAT and deployment schedule as not every sprint will make it into production after QA gives the OK to move into staging. There are pairing exercises that can happen once a developer is done an initial coding of a task to ensure that both a tester and an end user sign off on what was built. This is our third or fourth version of trying to integrate quality control into the project so it is still a work in progress that has evolved a few times over already.
Like a few other respondents have indicated, Testers should be involved from day one. In Sprint zero they should be involved in ensuring that the Stories the Product Owner is producing are testable (e.g. verifiable once coded) and "acceptable" (i.e. when you go though UAT). Once the Product Backlog is initially populated then the Testers can work on test cases for the Stories slated for the current Sprint, and once there is a product for them to test (Ideally somewhere in your first Sprint) then they can start testing.
If it sounds like there will never be anything to test for a few Sprints, you've got your stories wrong. The aim of a Sprint, even an early one, is to have a thin slice of the eventual system. Focus on "asprin" (i.e. if building a drug prescription system, how do you deliver testable functionality in 2-4 weeks? Build the ones for prescribing an asprin) and "tracer bullets" stories (ones which, when taken in combination touch all the risky parts of the architecture). You'll be amazed what you can hand over to test early on. If testers do end up with spare time, get them to pair program with the developers. It'll build relationships and mutual respect.
The benfits of this approach are many but primarily you test out a good deal of the internal people-processes of your development (handovers from requirements, to development, to test, and also the reverse) and secondarily the whole team (all three disciplines mentioned) sees the benefits of rapid feedback as a result of producing executable software.
It sounds impossible, but I've seen it work. Just make sure you don't bite off too big a chunk to begin with. Let yourselves ease into it and you'll be amazed.