DEFECT LIFE CYCLE
 |
Customer gives requirements –
developers are developing the s/w – testing team is writing test cases looking
at the requirements
Developer develops the product – test engineer
starts testing the product – he finds a defect – now the TE must send the
defect to the development team.
He prepares a defect report – and
sends a mail to the Development lead saying “bug open”.
Development lead looks at the mail
and at the bug – and by looking at the bug – he comes to know to which
development engineer developed that feature which had a bug – and sends the
defect report to that particular developer and says “bug assigned”.
The development engineer fixes the
bug – and sends a mail to the test engineer saying “bug fixed” – he also “cc
mail” to the development lead.
Now the TE takes the new build in
which the bug is fixed and re-tests it again – and if the bug is really fixed –
then sends a mail to the developer saying “bug closed” and also “cc mail” to
the development lead.
Every bug will have an unique
number.
If the defect is still there – it
will be sent back as “bug reopen”.
“REJECT”
BUG
Now, when the TE sends a defect
report – the Development Lead will look at it and reject the bug.
Bug is
rejected because,
1)
Misunderstanding of requirements
2) While
installing or configuring the product – wrongly configured or installed
the product and we found a bug in the product – send it to development team –
developer says “reject” because he looks at the defect report and comes to know
that the Testing team has not installed the product correctly.
3)
Referring to old requirements
Chances are there that Testing team may
be referring to old requirements while testing the product.
For ex, - in the 1st
build – developers have given the product for testing in which a sales link button is there, test
engineer does testing and reports any defects – in the 2nd build,
customer has given a requirement change to the development team where-in he has
asked them to remove the sales link
button – but the testing team is not aware of this change – so when the
development team gives the new build – obviously they would have removed the sales button – test engineer when he
looks at the feature – he is referring to old requirements which has sales link button – but in the new
requirements it has been removed which the test engineer is not aware of – so
he reports a bug – development team then reply back saying “refer to new
requirements in which sales link button has been removed”.
4) Because
of extra features
Consider the example and figure
shown in Pg 130.
Here, “help” button is a feature which has been developed by the
developer as an extra feature not part of he requirements – obviously when the
TE looks at the application and requirements, since Help feature is not there –
he reports it as a bug – the developer rejects it saying “it is an extra
feature and that he wont test it” – Test Engineer replies back by re-opening
the bug saying “update the requirements”
– updating the requirements requires lot of running around for the development
team, talking to the customer and all that – so he either removes it or talks
to the customer.
“DUPLICATE”
BUG
The TE finds a bug and sends a defect report
and also assigns the bug report a number – let’s say he sends a bug report
numbered as Bug 25 – now the developer replies back by saying “Bug 25 is
duplicate of Bug 10” i.e, another Test engineer has already sent that bug
earlier and it is being fixed.
Why do we
get duplicate bugs?
·
Because of common
features
– Lets say Test Engineer A and B are testing an application – Test Engineer A
and B to test their features must Login to the application using valid username
and password before they can go deep into the application and start testing
their features – A enters valid username and password and clicks on Login
button – it goes to a blank page – this is a bug – so A prepares a defect
report for this bug and sends it to developer – After sometime, B also tries to
login and finds the same bug – he also prepares a defect report and sends it to
developer – but developer sends back defect report of B saying “it is a
duplicate”.
·
B finds bug in A’s module – let’s say A and B are
testing their respective features – A is doing functional testing and
integration testing on all his features – Similarly B is also testing all his
features – In one scenario, B needs to go to A’s feature(say m12) and send data
to B’s feature(say m23) – when B clicks on m12, he finds a bug – although its
feature of A, he can still prepare a defect report and send it to developers –
and B sends the defect report to development team – now, A after testing all
his features comes to his feature m12 – he also finds the same defect and sends
a defect report – but the developers send it back saying “it is duplicate”.
How to
avoid duplicate bugs?
QA
|
|
|
|
|
|
|
|
|
|
Repository |
MS – Search
Let us consider Test Engineer A is
testing Sales module and finds a defect where-in – he enters all the
information and then when he clicks on Submit button, it is going to Blank page
– he prepares a defect report in which he explains the defect.
Now, before he sends it to developer
– A logs into a server named QA – there, he goes into a folder called “Defect
Repository” which has all the defect reports prepared by TE(s) and sent to
developers and clicks on MS-Search and gives Sales as the search keyword and
hits on search button – now, it starts looking inside the folder if there are
any defect report which has Sales in it – if the search results in No files –
then A will first copy-paste his defect report in the defect repository and
then send the defect report to the development team who then assign it to the
development engineer.
Now, after a few days – another TE B
finds the same defect and he also prepares a report – but before he sends it to
development team. He first logs onto QA and Defect Repository and searches for
Sales keyword – he finds the report which was earlier sent by TE A – so he
doesn’t send the defect to development lead.
But, always – the TE if he finds a
defect similar to his if the Search operation yields some results – before he
concludes that his defect has already been sent – he must first read the defect
report and ascertain whether it is the same bug which he has found – chances
are there that it might have the keyword Sales, but it is about some other
defect in Sales feature.
For ex, - TE A is testing password
text field – he finds that it is accepting 12 characters also (when requirement
says “it must accept only 10characters”) – that’s a bug. So he prepares a
defect report which has the keyword password and stores it in the defect
repository and also sends it to Dev team. Now, another TE B is also testing
password field – he finds there is a bug where-in password text field is also
accepting blank space character – before he prepares a report, he goes to the
defect repository and searches for Password keyword – there he finds the
earlier report sent by A – but that’s a different defect – he reads it and
realizes that it’s not the same defect which he has found – so, he prepares a
defect report regarding the blank space acceptance in the password text field –
and stores it in the defect repository and also sends it to the development
team.
CANNOT
BE FIXED
Chances are there – Test Engineer
finds a bug and sends it to Development Lead – development lead looks at the
bug and sends it back saying “cannot be fixed”.
Why does this happen? – Because,
·
Technology
itself is not supporting i.e, programming language we are using itself is not
having capability to solve the problem
·
Whenever
there is a bug in the root of the product. If it’s a minor bug, then
development lead says “cannot be fixed”. But, if it’s a critical bug, then
development lead cannot reject the bug
·
If
the cost of fixing the bug is more than the cost of the bug itself – cost of
the bug means loss incurred because of the bug.
POSTPONED
or Fixed in the next release
a) We find a bug during the
end of the release (could be major or minor but cannot be critical) –
developers won’t have time to fix the bug – such a bug will be postponed and
fixed in the next release and the bug status will be “open”
|
|
|
|
|
|
Developers have built the above
application – Test Engineers are testing the application and they find a bug –
now, the bug is sent to the development team.
Development team looks at the report
– and replies back to the testing team saying “they will not fix the bug for
now, as the customer is thinking of making some changes to that module or he
might as well ask for a requirement change where-in the module itself might
have to be removed” – so they won’t fix the bug for now – but if no requirement
change happens, then they will fix the bug.
c) If the bug is minor and
it is in feature exposed to internal users.
For ex, consider Citibank
employee using a s/w where-in he is making keeping track of accounts and
customer details – now, for example if the “Sort by name” feature is not working
– it’s ok because it’s a minor bug because the development team can always fix
the bug during the next release.
NOT
REPRODUCIBLE
|
|
|
|
DEFECT REPORT
Open the application in Mozilla FireFox
Let us consider the above
application – the TE is testing the above application in Mozilla Firefox– when
he clicks on SALES link button – it goes to a blank page – this is a bug – the
TE prepares a defect report and sends to the development team.
 |
Open the application in Internet
Explorer
The development lead looks at the
defect report and opens the application in Internet Explorer – when he clicks
on Sales link button, Sales page opens and the application works perfectly.
Then, what was the problem? – the
answer is, the TE didn’t specify in the defect report in which browser to open
the application – so, the application didn’t work for the TE in Mozilla FireFox
– but, when the developer opened it in Internet Explorer, the application
worked totally fine.
Since the developer is not able to
see any defect – he replies back by saying “I am not able to find the defect”
or “I am not able to reproduce the defect”.
Why we get
“not reproducible” defects?
1. Because of platform
mismatch
2. Because of improper
defect report
3. Because of data mismatch
4. Because of build mismatch
5. Because of inconsistent
defects
1) Because
of platform mismatch
Ø
Because
of OS mismatch
Ø
Because
of browser mismatch
Ø
Because
of ‘version of browser’ mismatch
Ø
Because
of ‘settings of version’ mismatch
2) Because
of improper defect report
Example to explain this - take
figure and explanation in Pg 136 and 137. Here, the TE must have specified to
open the application in Mozilla Firefox
Let us consider another example shown below to
explain “improper defect report”
 |
|
|||
DEFECT
REPORT
In the above application – TE is
testing the above application – when he enters all the user details and selects
checkbox and clicks on Submit button – it goes to blank page – but, when the TE
does not select the checkbox and clicks on Submit button, a new user account is
created, i.e, it works fine. The TE prepares a defect report and sends it to
development lead.
The Dev team looks at the report(as
shown in box) – and tests the application – it is working absolutely fine.
Then what happened? – Observe the
defect report sent by the testing team – nowhere as he mentioned to “select
checkbox and click on submit button” – thus because of the improper defect
report – the development team is not able to reproduce the defect.
3) Because
of data mismatch
Consider the example shown below,
 |
Login as
ABC and click on Inbox in homepage of ABC
The above email application has been
developed – the TE starts testing the above application – after testing and
testing the above application using username ABC and valid password – there are
almost 101 mails in ABC’s inbox – now, again when the TE opens the application
and goes to test the application – no mails are being displayed when he clicks
in Inbox and instead he gets a blank page. So – the defect he has found is that
– when the number of mails in Inbox exceeds 100 mails , he gets a blank page –
so he prepares a defect report as shown in the box below,
|
The development team looks at the
defect report – he logs in to the application using username XYZ and clicks on
Inbox – no blank page is displayed and the Inbox is working fine.
What happened here? – The TE should
have mentioned in the defect report as “Login
as ABC” or “Login to any mailbox
which has more than 100mails”. Thus, the developer does not find any
defects and thus he mentions it as “not reproducible defect”.
4) Because
of build mismatch
 |
Developer develops build 1 and gives
it for testing – in Build 1, there are 2 defects – bug 1 and bug 2 – TE while
testing finds bug1 and reports it to development team – development team when
they fix bug1, bug2 automatically gets fixed – development team is developing
build 2 and gives it to testing team for testing – testing team just when it
finishes testing build1 – he finds Bug 2 and reports it to development team –
the development team however do not find any defect as they are testing build 2
– so they report back saying “bug is not reproducible”.
5) Because
of inconsistent defects
To explain this, let us consider an
example : When TE is testing an email application – he composes mail in User A
and sends it to User B – he then logs out from user A and logs into User B and
checks B’s inbox – but no mail is there in the Inbox – thus he finds a defect –
now, just to confirm the defect before sending a defect report – the TE again
logs into User A and sends a mail to User B – he then logs out from User A and
logs into User B and checks B’s inbox – but, this time the mail is there in B’s
inbox!!
Thus, we don’t know when the defect
comes and when the feature works fine. This is called inconsistent defect.
REQUEST FOR
ENHANCEMENT
(RFE)
Test engineer finds a bug and sends
it to development team – when development team look at the report sent by the
TE – They know it’s a bug, but they say its not a bug because its not part of
the requirements.
Let us consider the example shown below,
In the above example, the TE is testing the above fields. After he enters all
the data into the required fields, he realizes that he needs to clear all the
data – but there is no CLEAR button
– and he has to manually clear all the fields. He reports this as a bug – but
the development team will look at the defect report and see that clear button is not mentioned in the
requirements – so they don’t reject or close the bug – instead they reply back
to TE saying that it is RFE.
Developers agree that it is a bug –
but will say that the customer have not specified these details and hence will
give it a name as RFE.
If a defect is known as RFE, then we
can bill the customer. Otherwise, development team needs to fix the bug
Development team always have a
tendency to call a defect as RFE, so a TE needs to check the justification
given by development team – if it is valid, then he will accept it as a RFE –
but if it is not, then he will respond to the development team with proper
justification.
|
DEFECT REPORT
Defect ID – it is an unique number
given to the defect
Test Case Name – whenever we find a
defect, we send the defect report and not the test case to the developer. For
defect tracking, we only track the defect report and not the test case. Test
case is only for reference for the TE. We always only send the defect report
whenever we catch a bug.
When
we are doing ad-hoc testing – no test case is written for ad-hoc testing
because they are “out of the box” testing scenarios – if we find a critical bug
– then we convert that ad-hoc scenario into a test case and send it to
development team.
Given below is – how a defect report
looks like – it is a MS – WORD file.
 |
||||
|
||||
Defect
Evaluation Team / Bug Counseling Team
 |
|||||
 |
|||||
 |
|||||
 |
|||||
Let us consider the above figure to
explain about Bug review meeting / Bug
Counseling
The Defect Evaluation Team (DET)
consists of Development Lead, Development Manager, Senior Development Engineers
and Test Lead.
They will have a group mail id – say
– xyz_det@abc.com – Now, whenever TE finds
a bug, he sends it to DET – In DET, only the Development Lead (DL) looks into
the bug report and assigns the bug – although the mail goes to everybody in the
DET, only DL will check it and assign the bug. This process continues from
Monday morning upto Friday afternoon – by this time about 100bugs have been
reported by TE(s) – out of these 100 bugs, 60 have been assigned and the
remaining 40 are pending.
Now, on Friday evening – DET team
will have a meeting where they discuss about the pending bugs and whether to
assign the bugs or reject the bugs – but for everything, they will have to give
proper justification.
This meeting is called Bug Review
Meeting / Bug Counseling. By the end of the meeting (i.e, 1st cycle)
– the status of all the pending bugs will changed from pending to other
status(es).
Manual
Automation
 |
 |
DEFECTS
DEFECT TRACKING
 |
 |
Manual
Automation
How to
track a defect manually?
1) Find a defect
2) Ensure that it is not
duplicate (i.e, verify it in Defect Repository)
3) Prepare defect report
4) Store it in defect
repository
5) Send it to development
team
6) Manage defect life cycle
(i.e, keep updating the status)
Tracking
of defects using Automation (i.e, Defect Tracking Tool)
The various tools available are,
®
Bugzilla
®
Mantis
®
Rational
Clear Quest
®
TeleLogic
®
Bug_track
®
QC
– Quality Center – it is a test management tool – a part of it is used to track
the defects.
There are 2 categories of tools,
1) Service based – In service based
companies – they have many projects of different clients – each project will
have a different defect tracking tool
2) Product based – in product based
companies – they use only one defect tracking tool.
Now, let us see how the defect tracking tool works,
No comments:
Post a Comment