Advertisements
Archive | Agile Testing RSS for this section

Build Pass when Tests Fail

It happens to all of us. We are used to doing a process in a particular way for years and never think of other ways of doing it. Then someday someone says something. That serves as an eye opener and we start seeing other ways of doing the same thing.

This happened to us for our rule: “Break a build if a single unit test fails”

Sounds very simple and rational. We have this rule for may be 10 years and I have repeated this over and over in all new projects that we took up in these years. Our structure follows the Plan A as shown below i.e. to run tests as part of the build process and if a single test fails, build fails.

What changed in year 2017 was a quest to find ways to release faster, you know the DevOps like stuff. So we started to look at the reasons for the build to be broken. We build our source 4-6 times a day, so we had enough data to look into. One of the reasons was always a failing test.

Now we thought, as you must be thinking by now, that it is a good thing. We should not ship a build for which a given test breaks. But our data (and wisdom) suggested that failing tests were in following three categories:

  • The underlying code changed but test code was not changed to reflect this. Thus test fails but code doesn’t fail.
  • The test is flaky (expect a full blog post on what flaky tests are and what we are doing about them). For now, flaky test is that passes and fails with same code on different occasions.
  • The test genuinely fails i.e. the feature is indeed broken.

Now 1 and 2 are important and Developer who wrote the test need to pay attention. But does this stop the build to be used? Of course not.

3 is a serious issue but with the notions of ‘Testing in Production’ combined with the fact that fix is just few hours away, we figured out a new rule as shown in Plan B.

Yes, when a build fails due to a failing test, we actually report bugs for failing tests and declare the build as Pass. Thus the wheel keeps rolling and if it needs to be stopped or rolled back, it can be.

A few weeks into this strategy while all looked good, it happened what was always feared to be happening. A build in which 20% of our tests were failing was declared pass and our bug tracker saw around 100 new bugs that night. Let’s call that spamming.

That raised our understanding to move from binary (fail or pass) to a bit fuzzy implementation. We came with a new rule that if 10 or more tests (where 10 is an arbitrary number) fail, we’ll declare the build as failed. Otherwise we’ll follow the same path. So this is now our current plan called Plan C.

I know what you are thinking that a single important test could be much important than 10 relatively unimportant tests. But we don’t have a way to weigh tests for now. When we have that, we can even improve our plans.

Does this sound familiar to your situation? What is your criteria of a passed build in context of test results?

Advertisements

How Code Reviews can improve Software Quality?

The very idea that quality can be added later is flawed and quality must be baked in.

Speaking of baking, it is not possible for the inspector who tastes a finished cake to make it any better. That person would just tell you if the cake is good or bad. And if it’s bad, there is nothing you can do about it. But if you want the cake to be always good, you will take good consideration on the ingredients that will go in, you’ll consider hiring a professional baker to do the job, you’d like to inspect it in intervals like how the dough is or opening the oven to see if cake is in good shape.

Similarly, you cannot have Testers jump late into the development cycle and improve the quality. A Tester can only tell you if the Software is good or bad. And if it’s bad, you’ll have to go many cycles of correction before it goes reasonably good. So, if you want your Software to be in good shape, you’ll take care of the technology and tools you use, you’ll consider training or hiring professionals to do the job right, you’d like to review the code time to time to see if it has quality baked into it.

Though all of the above topics are of interest to anyone who is serious about producing quality software, let’s focus on Code Reviews today and see how it can help in improving the Software Quality.

Catching logical errors:

All of us know that humans make errors but all of us also believe that those humans don’t include us. It might be true that some people make more mistakes than others but no single person is error free. So as a matter of fact, when you show your code to a peer to review it, or you explain it yourself to a peer through a walkthrough, you start catching errors. Some of these are obvious logical errors which got slipped as you wrote that code. Don’t feel ashamed as this happens and we should be happy that this error was caught much early in the cycle.

Catching Performance issues:

Writing algorithms that solve complex problems is not an easy job and at times in an attempt to do that, you can take a path that is lengthy. Well, lengthy in terms of how many CPU cycles you’ll need to complete them which can result in slow performance. There are also cases where unnecessary data structures are involved which cause more memory to be consumed than the minimum required one. Your peers can help you find such errors in the code. You can also have some expert on this subject in your team to do this type of review for all the code that is being written.

In one of the teams in which I worked, we had a person (let’s call him Ahmed) who was our performance expert. All team members when confronted with the task to write best performant code would consult Ahmed after writing the code. And if Ahmed was happy about it, it was always a quality code.

(image source: https://www.dreamstime.com/stock-illustration-code-review-magnifying-glass-focussed-source-concept-to-reviewing-image55133888 )

Memory leaks and Security considerations:

Most developers are capable of writing code which is free of logical errors and is usually well performing, but there are other things hidden like the bugs who only come out when there is dark and no one is around. Yes, I’m talking about writing secure code.

These aspects of code can again be reviewed by Developers who are experts in this domain. The good thing is that some of this expertise is now being passed on to tools. For example, you can use built-in tools in your IDE or can get many free tools that can help you find memory leaks in your code.

Similarly, if your code is good in terms of all requirements but is not safe and is vulnerable to various type of attacks, we cannot call it Quality code. Before some hacker can detect those vulnerabilities, how wonderful it would be knowing these so early in your development lifecycle.

Knowledge of the tribe:

The collective knowledge of your development team is what matters more than the individual member of your team. In fact, the sum is always greater than the actual sum. Let me explain this through an example.

Let’s assume that you three members in your development team named Ali, Sara and Jamshed. K represents the knowledge possessed and the brackets are meant so as who possess that knowledge. Given the above, the below is always true:

K (all team) > K (Ali) + K (Sara) + K(Jamshed)

So, the real job of management is to create an environment in which knowledge of one person is passed on in a way that it becomes a knowledge of the whole tribe. Code Reviews play a big role in achieving this.

I hope that by now you are convinced that you need to do Code Reviews. So maybe I can help you further on the how part of it where you can ask these questions:

How often should Code Review happen?

There are many school of thoughts on this. Some believe that all commits should be reviewed such as if it is not reviewed, it will not be in the production code. Others think that achieving this “all code must be reviewed” policy takes lot of effort and you should be selective in the process.

If you pick the first path, there are tools that can help you such that no code goes unreviewed. And if you like the second approach more, you’ll have to decide what is the selection criteria and who makes the decision if a particular piece of code ought to be reviewed.

Who should do the Code Review?

Just like the above, there are all and selective approach. Some teams like each member of the team to take the Reviewer role once in a while. In my opinion it helps to give this power to all rather than few. If only some team members review the code of other team members e.g. senior folks reviewing junior folks code, it will create friction in the team.

The exception is obviously special cases where you have Security and/or Performance experts in your team who’ll do more reviews than other members.

How long each review should take?

Though this largely depends on how big is the piece of code being reviewed but leaving this as an open time slot can both slow down the process along with the code author feeling being questioned for too long. You can set sort of “30 minutes or one hour rule” that each small code review should take no more than 30 minutes and each big review should take no more than an hour. These numbers are arbitrary and you have all the powers to set a limit of your own.

What the Reviewer should check?

Typically creating a checklist will help you which can have questions like the below:

  • Are there any errors?
  • Is code human readable?
  • Is there good commenting?
  • Are there tests written against this code?
  • The coding guidelines have been followed

These questions vary from organization to organization and team to team. So, you should feel free to start with your own list and then grow this list as you see other cases.

All the best on your journey of making this world a better place by producing Quality Software.

 

Note: The above article was printed in Flash newsletter which is being reproduced for blog’s audience.

What’s this DevOps thingy?

DevOps is a big thing these days and it is hot. Lot of people are talking about it and as we are going through this journey of adapting DevOps culture, we are learning it.

In recent month or so, I have given some introduction to fellow testers at my organization and shared thoughts at couple of Industry sessions. So I thought it would be good idea to share my slides here.

For further study, I’d recommend reading following content:

All the best with your DevOps journey!

Agile Conference Pakistan 2017 – My learning

The signs of a growing IT industry in Pakistan are evident with the number of events that happen each year. One such event is the annual conference on Agile which saw it’s 4th edition on October 21, 2017. The Agile Conference Pakistan 2017 hosted by Pakistan Agile Development Society hosted over 300 professionals joining from around the country. ClickChain team from Bahawalpur lead by Muneeb Ali, Contour Software team from Karachi under Owais Ashraf leadership, and half the panelists from Lahore provided the best possible mixture of top IT talent at display during the day. The theme this year was “SCRUM in Pakistan”.

I have provided detailed reports for the 2015 and 2016 editions but this time rather than a report style, I’ll share some of my learning from the day.

Do use it at Home:

For many commercials and TV programs with stunts, they have a message like “Don’t do it Home”. Naveed Khawaja who is a UK based Agile Trainer and Coach (and adjusted his schedule to be able to join us) mentioned that because he loves Kanban, he does all his personal work through it. He shared examples like:

  • His 6 year old daughter using a Kanban board.
  • His family using a Ramadan Kanban board.
  • His (poor) carpenter being Kanban-ified for some renovation work.

The idea that I got is that if you like something, you can try it for both your professional and personal life. Interestingly, after I attended the 7 Habits training from FranklinCovey few months ago, I practiced the learning both at office and at home.

(more photos at: https://www.facebook.com/bendaoods/)

Purpose is more important than Terminology:

Sumara Farooq in her talk suggested that when SCRUM or any new initiative is implemented in an organization for the first time, the message can be simplified by focusing on the purpose or intent of doing something rather than using the terms. For example:

  • A Storyboard is actually there to provide visibility.
  • A Daily SCRUM is actually needed for better collaboration

This indeed can change the way you communicate with the real folks who’ll embrace the change if they like the purpose rather than saying that “we need SCRUM from tomorrow”.

Performance = Skill x Will:

Mohsin Lodhi was fantastic in his talk as usual and threw in lot of interesting ideas for anyone who manages people. His main theme was Servant Leadership such that what it is and how to achieve that. As he was setting up the idea, he shared the above formula which I really liked. I have seen many teams under performing because either they lack Skill or they are not Willing to do it. I learned that both of these are important and any leader should work on enhancing Skill of the team and buying their Will to perform.

Selling your idea is the main thing:

The panel discussion which had Naeem Iqbal, Naveed Khawaja, Faisal Tajammul, Shaima Niaz and was moderated by me saw a flurry of question from the audience. The question ranged from difficulties in SCRUM transformation, lack of top management support, collaboration issues, fear of failures etc.

A common theme that I saw in the questions was that you know why an idea is worth (say SCRUM) but you are unable to convince others. I call this a “selling problem” and as someone rightly said:

Selling is not part of the game, it is the game.

Unfortunately all of us, the IT guys think that if a solution is technically good, people would love it and buy it immediately. But buying something is more psychological than technical and we need to learn the art of selling. I am learning it too but have found one thing: if you target on the need of the buyer, you might be able to sell something. For example in case of SCRUM, consider selling cost saving and faster results to management and consider selling self-organization and better prioritization to your team members.

There were lot of amazing people that I met over the conferences and many interesting discussions (including Khurram Ali‘s talk on User Stories, Muhammad Ibrahim‘s research presentation on Scrum and XP) that I saw which I’ll try to cover in future posts. Before I go, want a special mention of Faiza Yousuf and Noorjehan Arif who came from Karachi for the conference and ran our twitter campaign. And yes, #AgilePK and #ACP2017 were the top trends on that day.

Thanks to all members of the organizing team which orbit around Naveed Ramzan, who worked hard to made it a memorable day. And Anum Zaib lifted the level of event host considerable from last year (last year it was me :))

What did you learn from the ACP2017 if you were there? And if you were not there, what else do you want me to cover about this event?

Providing valuable information as Testers

The best answers to “what is the job of testing?” always have a key component suggesting that it is “to provide information”.

Jerry Weinberg defines it as following in Perfect Software book:

Testers job is to provide information for the product under test

Johanna Rothman defines it as following in in the Preface to More Agile Testing book:

They provide information about product under test, to expose risks for the team.

Now if you have been provided information any time in your recent past, you can easily recall that some of it was valuable and some of it was not. Even some information which was provided to you was false, misleading, hiding the facts or inflicted with similar diseases. So how do we, the testers, can provide information which is always always valuable to the person who is receiving information?

I believe the answer is in two parts: a) How do we collect information? and b) How do we process and present the information?

May be I am bit vague above, so let me spend some time to explain myself.

To provide information, you need to have all the information available to you collected from different sources like test results, design documents, release schedules etc. And then you need to process this information for the given need like the occasion, the person or timing of the information being provided. Like providing specific filters to view information from a certain angle.

Let’s dig a bit deeper.

How do we collect information?

Some testers are overly relying on only the results of testing. They don’t like to discuss with Development about how those tests are designed or how the results are interpreted rather they believe that results of testing are the only source of truth.

Then there are others who love documents. All of the information they have is primarily coming from a document e.g. Design document, Project Plan, Release Plan, or some sort of change tracking system like TFS. For them, those documents are the only source of truth.

And there are the third group who like verbal communication. They discuss each and every type of test in daily SCRUM calls and get feedback on Iteration Review results. They discuss release plans with the management rather than looking at what the document suggested at the start. They are of the view that people involved in the team are the only source of truth.

Not many people know that if all the information you have is coming from a single source, you don’t have a complete view. Consider political arguments that happen in your work environment between people who watch/follow a particular news channel.

So if as a Tester, you want to improve your skill as a better information provider, the first step is to widen your input sources. Test Results, Documents, People all matter. And even there are other sources which you know better than me.

How do we process and present information?

So when you are asked “How was your weekend?”, your answer largely depends on who is asking and in what circumstances. Your answer will vary in details and emotions for your jogging partner compared to your colleague; your best friend compared to someone you are meeting for the first time at a business meeting.

Similarly when you are asked “How is the testing going?” or “Are we release ready?”, if you are giving the same answer to every person in every possible settings, it shows that you are not filtering the information you are having and passing it on as it is. Apply different views to the same information given the context and your answers will contain valuable information for the receivers.

What tips you have to improve information gathering and processing steps?

Five years of blogging

KnowledgeTester turned 5 few days ago!

And as I discussed the journey in first, second, third and fourth year, my first happiness is to see that it is still going. I am, just like many of you, type of a person who “has a habit of starting projects and not completing them”. Though the project of blogging never finishes but continuing it for 5 years certainly can be used as a yardstick to call it mature.

(the original image is here: http://learnchineseinyunnan.com/news/news-wire/2015/08/20/2015-the-fifth-anniversary-of-thanks-giving/ )

The biggest achievement of fifth year is not a personal one but a collective one. And I sincerely hope you know that collective success has more savor than a personal one.  And that was the first ever testing conference in Pakistan that we saw happening in April 2017 in a great way. My report on it “Pakistan Software Quality Conference 2017” has been the most viewed post of the year and I’m so happy to have achieved this through all friends specially the support of Pakistan Software Testing Board.

Interestingly that was my top most priority when I recapped fourth year. Now you can easily guess that this year’s priority is to make PSQC a regular thing in our Industry and we are working for PSQC’18 which will happen in February 2018 in Lahore Insha Allah (God Willing). Though we went into a momentary post conference enjoyment mode, where we made little progress in recent months but we are now getting back with two meetups coming soon in the remaining months of 2017.

On the viewership increase, I am seeing a 20% increase in per post views which may not seem great compared to previous years. But you know as the baseline increases, it becomes hard to keep increasing at same pace e.g. think how salary raise in % is tougher as your salary gets higher. But net effect is that views per post are almost at 1,000 now (967 to be precise) in last year. Yes, I can’t believe that this viewership is for the stuff that I’m writing but it’s true 😊

During the last two years or so, due to the community building activities KnowledgeTester was low on providing training to Testers. The good news is that I’m setting this as my second goal for the sixth year i.e. to hold Training along with continuing community activities. Wish me luck.

That’s about it for now and I’m thankful to each of you who contributed towards this success. And I’m sure you’ll be with me in next years of this journey.

Do you have any suggestions on what kind of training we should be offering this year?

3 Tips for near-perfect Test Planning

Every plan is perfect until the execution happens. As a losing team’s coach once said: “We had a perfect plan on the paper. The only problem is that the match was played on the ground and not on the paper”.

So should we stop striving for better planning? Of course not. Rather we should learn from each execution so that our next plan is better than the previous ones. Taking a hint from this thinking, let me share 3 tips based upon execution failures to improve your test planning.

I have selected the tips based upon 3 layers of what I’m calling Agile test planning. Such that as a Tester, you are required to do “Release Planning” or asked to do an “Iteration Planning” or asked to do an “Individual Tester Planning”.

Let’s take an example of “Release Planning” first.

Many years ago my Employer decided to start offering solutions in the Mobile world. Before that we were only in the Desktop world. Our first solution was for Android and then we worked on a Platform that can produce Apps in Android, iOS or Windows. A year or so after this has happened, I was asked to do some test planning for a release meant for a Client in China. I ‘assumed’ that they’ll need mobile solution as it was our top priority then. So my plan started with Android and then iOS testing on Simulators and Devices.

The plan looked great. Our first iteration started and by that time, I had no interaction with the Client. Client started reporting a lot of bugs and all of them were reported in Windows Desktop environment. I learned that this Client is only interested in good old Windows Desktop solution and doesn’t care about mobile.

The lesson which I learned is my tip # 1 for this blog:

Always ask question rather than assuming things

There is an excellent resource on Ministry of Testing that gives you a good list of what questions to ask for test planning. Michael Bolton also has a list of questions.

Let’s now move to “Iteration Planning”.

In a project that I was part as the lead tester, the Product Owner complained that the number of new bugs reported towards the end of Iteration is far greater than the ones reported at the start. My first response was that as more things get ready for testing towards the end, more testing is done and hence more defects were found.

Then I gave it a deep thought by looking at the history of tests we were executed. One pattern was there that certain test plans produced higher number of defects (following the heuristic that Software Bugs are like Real bugs) and the test execution order was left on the tester to pick. Tester knew that this test plan will consume lot of time and energy, so they left these to the end.

The rule that we made afterwards is my tip # 2 for this blog:

Prioritize test execution based upon which tests will fail first

This article can help you get started on Risk Based Testing.

Let’s look at the last one called “Individual tester planning”.

When I was young and raw, I had very ambitious plans for myself assuming that I’ll get all things in my favor. I planned it 8 hours a day with no distractions. Eventually I couldn’t always finish the stuff that I committed at the start of Iteration.

With age and lot of experiences, I have become more realistic about my estimates and always consider it as a 6 hours day. I also account for unknown distractions like network outage, random meeting popping up, sick days and what not. It’s not that I give a pessimistic by including all those factors but I do keep margin for some of them.

The lesson that I learned is my tip # 3 for this blog:

A plan that doesn’t incorporate change is not a plan

What tips you can offer from your execution failures to improve test planning?