Subscribe by Email

Thursday, March 19, 2015

Trying to solve defects where the flow is not repetitive .. (contd..)

In the previous post (Trying to solve defects where the flow is not repetitive ..), we talked about defects that are not easily reproducible, and some points around that. In this post, we will talk about such defects with more points. Handling defects that are not reproducible is one of the biggest pain points of the development cycle, since you know that even though the defect is not easily reproducible, it will come back to bite you at some point or the other. So, you cannot ignore those defects, although after trying to reproduce a defect like this for hours, it is tempting to just forget those defects. It needs a strong policy to ensure that the team does not overlook such defects.

- The team needs to have a policy about handling defects that ensures that the team does not overlook some of these defects. The policy needs to define how the defect needs to be handled, how it is tracked, how many times does the team needs to re-test those scenarios, whether such defects need to be handed over to other testing personnel, and so on. These policies need to be ensured as a part of the system, so that people know exactly how to handle such defects.
- These kind of non-reproducible defects do not follow any easy pattern, but from time to time, groups of developers and testers do figure out steps on how to figure out how to reproduce some of these defects. I remember specific cases where a developer wrote an app that would be run along with the application during the testing process, and helped in around 10% of the cases of non-reproducible defects, and that was an amazing feat, since it helped save a lot of time and helped in resolving defects that were otherwise difficult to fix.
- Statistics for these kind of defects is very important, and during the development process rather than at the end. What these statistics for non-reproducible defects do is to determine for the team management about how many these kind of defects are being found, as well as how many are being resolved. Every team needs to specify a number beyond which action needs to be taken. For example, during a specific 3 week period, we found that there was a spike in the number of non-reproducible defects that were being logged. This in turn resulted in more time being spent by the testers on trying to reproducible and hence resolve these defects. A solution needed to be found for this - and we spent some more time on focusing superior development effort on these defects. Eventually, this extra effort helped, since it was a new version of an external component that was causing problems to the application and resulting in many such non-reproducible defects.

As a result, even though it is difficult to visualize handling such non-reproducible defects, they are a reality and teams need to have a process of handling such defects.

Sunday, March 15, 2015

Trying to solve defects where the flow is not repetitive ..

If the title of the post did not make sense, then let me explain. What I refer to is the case where a defect has been logged (let's not discuss the severity of the defect right away), but the flow to replicate the defect is not easy. Anyhow, it is easy when a defect can be easily replicated, when the steps are easy. However, for anybody who has gone through the whole process of defect detection, analysis and resolution, finding an easy method of replicating the bug steps can almost be counted as lucky. There can be numerous cases where the steps for replicating a defect is not so easy.
Consider the case where a defect is caused due to a particular variable loading a specific value. This value has not been loaded during the current workflow, but gets loaded during a step in an unrelated workflow. Now, depending on whether the user has gone through this unrelated workflow previously, the defect will show up or will not show up. One might think that the person would be able to replicate this defect easily, but this is easy to think post-facto once the defect process has been understood. For the tester, there is no way of easily knowing that the defect is getting caused due to the operation of an unrelated workflow; it is only the skilled tester who tries the workflow inside the product multiple times who can try and figure that the defect is happening only in those cases where the unrelated workflow has been used. These kind of stories are common (not the majority of cases, but common enough that testers and developers need to be aware of these issues).
What does one do in these kind of cases ? Well, such cases are handled different by different teams, but let me suggest a few points for these:
- It is important to understand that every defect is important, and whether the defect is fixed or left unresolved needs to be decided in a determined manner. For a customer, the defect can be problematic, especially when the person used the unrelated workflow and wonders why the team left such a defect not fixed.
- Another variable that decides what needs to be done is the severity of the defect. For defects that were severe, it is not easy to decide to close the defect as unresolved. For defects that are minor, the bug decision committees would need to decide whether these minor defects need to be fixed, or can be deferred; that the team was comfortable with the minor inconvenience that such defects would cause to the customers.
- It is not easy for the developers to fix defects that the testing team cannot easily replicate, but that does not mean that they should not be fixed. There will be a larger number of customers than the testing cases within the development team, and one can be sure that some of the defects that could not be replicated will show up and cause inconvenience to the customer,
- There are methods that the development team can use to try and make it easier to find a solution to the defect. For example, it might be possible for the testing team to use special development versions of the software where the values of key variables can be recorded during the testing process, and this would help in trying to figure out exactly what is going on; and there are other development steps that can be done to try and get more detail on what exactly is causing these defects.
- Teams can setup processes on how to treat such defects; letting the tester spend some adhoc time on these defects; or give these defects to a different tester in order to get a fresh set of eyes; or ensure that the developer has some time for these defects in order to try and get some sort of solutions for atleast some of these defects.

Read more about this in the next post (Trying to solve defects where the flow is not repetitive .. (contd..))

Monday, February 23, 2015

Tracking platform usage and making decisions based on that

Even in today's data based world, if you are an analytics expert, you can't expect to be totally popular, or that people will welcome you with hands outstretched. There are many aspects of a product development cycle that would benefit from integration with a data analytics cycle - generating the questions, collecting the data, and the extremely important task of generating the output information and conclusions (and a wrong conclusion at this point can have negative implications for the product - putting in effort in wrong or unimportant areas). However, consider the case where there is support for using analytics and there are resources dedicated for ensuring that requisite changes to the product can be made (based on the analytics output).
One major area where the team needs to gather information is on the area of platform support. Platform could mean the Operating System (MacOS, Windows, others), Versions (Windows XP, Windows 7/8, etc) as well as browsers support (versions of Chrome, Internet Explorer, Firefox, Opera, etc). These can make a lot of different in terms of the amount of effort required for product development. For example, on different versions of Windows, the versions of systems files are different and it is a pain to support these different versions. There can be small changes to the functionality of the system based on these files, and in many cases, the product installer would actually detect the Windows version and install the required versions of these systems files. If you could find out that the number of people using these different versions, and find out that one of these versions is being used by a small number of consumers, then the decision to drop that particular operating system may be taken.
So, this is one of the areas in which analytics can work. The simplest way to do that is to make it part of the functionality of the product, where the product, on the user's computer, dials home and provides information such as the platform on which the product has been installed. Once this kind of data has been collected and a proper analysis is done, then the product team can look at this information and factor that into the feature list for the next version of the product. The team will need to decide on the time period for which the data would be captured, and also some of the benchmarks which will decide whether data from the analytics effort needed to be used for making decision (for example, the data seems to be such that people feel that the data is widely different from public perception).
However, this is also dependent on how much the team can depend on this data and the analysis; after all, even a small variation during the analysis can result in information that has levels of inaccuracies in it. But, it is necessary that the team spends effort in the analytics effort, since the payoff from using accurate data analysis and interpretation is very high.

Wednesday, February 11, 2015

Trying to get non-responsive members of the team be more schedule sensitive

We know this problem, it happens all the time. You have different members of the team, some more disciplined and some less disciplined. Actually discipline is the wrong word. When you have creative members of the team, or team members who are attached to multiple projects, then there can be problems with respect to scheduling of their deliverables. In the case of team members such as User Interface Designers or Visual Experts, or Visual Designers, they typically do not move to the same beat as that of the rest of the project teams, such as Engineers or Testing Engineers.
This can be problematic for the rest of the team, since schedules are interlinked to each other. For example, the User Interface Designer would prepare design specifications that are used by the team members to discuss and finalize the feature workflow and the technical architecture. These are then developed and passed onto the testing team which does the testing and then releases the feature. However, if the initial design does not come in time, then the rest of the schedule will get impacted.
One of the problems that I have experienced with User Designers or similar creative people is that they do not work in pieces; they would like to look at the overall workflow for the product and then release a completed design. But the team does not work like this, it would like workflow designs feature by feature, so that the work can be done feature by feature (and it makes logical sense).
Another option that could have been postulated is that the Workflow Designer could have a period of 2-3 months before the start of the cycle, so that the Designer gets enough time to make the design. This seems logical, but there are problems in this. The Workflow Designer does not work entirely on his / her own, but needs to work with the Product Manager and the team members (the team members  are involved so that the team could figure out the technical cost of doing the Workflow designs; some of these workflows may take more time and effort than other workflows and the contribution of the technical team in figuring out these is critical. This process can be iterative).
So how do you work out trying to get such more creative members as part of the process?
- First and foremost, it is necessary to ensure that you do not make the assumption that these resources understand the critical nature of meeting their schedule deliveries. It would be needed to spend much more time with these people and form a detailed plan for deliveries, doing this discussion multiple times till an understanding has been formed.
- In my experience, it was also necessary to have 2 dates in the schedule with a few dates gap between the 2 dates. It was necessary to push the delivery to happen for the first date, but there was also the understanding that the delivery happening on the 2nd date would also work fine without threatening the schedule.
- It was also realized that there was the need for a regular reminder along with checking about state of progress and updating the rest of the team on such progress. So, the Project Manager had setup a weekly meeting with the workflow designer to discuss the state of progress and the deliverable, and figure out alternatives if there was a delay.

Thursday, February 5, 2015

Emergency defect fixing: Giving local fixes for quick verification

During the process of defect fixing and verification, there is a standard process whereby a build process is created and defect fixes checked into this build. What this does is to ensure that every day (the builds typically come every day) the defects which were fixed the previous day are available for testing in a proper installer which can be used by the testing team similar to the product that is available to the customer.
This process works pretty well, as long as everybody involved knows the process well, and there are people in the process who have responsibilities for the different parts of the process (for example, somebody who ensures that the build stability systems are in place, others who do a quick smoke testing to ensure that the build is usable), and so on.
However, such a system cannot protect against a case where a defect has not been fixed, either fully, or partially. In the normal case of operations, it is normal to have defects that are not fixed and are rejected by the testing team, or some part of the defect not being fixed well, and a new defect being filed for the same.
When does this process not work ? Consider the case where the product development process is nearing the end of the schedule. In such a case, the defects that are to be fixed are restricted, and only those defects that are being allocated for fixing are passed onto the development team and the list given to the testing team for verification. However, the cost of a failed defect fix can be fairly high.
A defect fix that has failed would mean that the build for that day is not ready for use, and this can be very expensive.
When such parts of the schedules have been reached, there is the need for much closer interaction between the specific developer and the tester(s) for that defect fix. When the developer has made the fix, he/she would work with the tester and provide the fix in a local build, made on the developer's machine, which could be quickly tested by the tester to ensure that the fix has been made to the satisfaction of both the developer and the tester. This goes a long way to help that the build that comes the next day is usable and important fixes are not failing.
There are some problems that can still happen in this process. The local build may not be incorporating changes made by the other developers, and this can cause a dependency problem that may still cause the defect fix to fail. However, the chance of this happening is low (or can be monitored by the developer to reduce the failure rate) and goes a long way to ensure that the development process has reduced risk near the end of the cycle.
However, this requires close collaboration between the tester and the developer, and is not really required to be done in the regular part of the development cycle, since there is an overhead involved in the process.

Sunday, October 19, 2014

How is Agile Unified Process (AUP) different from Rational Unified Process (RUP) ?

Agile Unified Process uses Agile Modelling, which is used for describing the modelling and documentation process based on agile disciplines. The Rational Unified Process (RUP) is also an agile methodology. Here we highlight the differences between the two. Both of these processes are divided into disciplines or workflows, which are carried out in iterations. Agile Unified Process (AUP) is derived from RUP, and we can say it is a simpler version of it. There are 3 disciplines followed by RUP i.e., Modelling, Requirements and Analysis and Design. The AUP being the simple version combines all of them into one discipline. It is because of this that it is easy for the RUP teams to migrate to AUP if they want to. Thus RUP is very flexible and can be merged with agile modeling practices.

> Active stakeholder participation: In RUP projects, stakeholders including customers, users, managers and operators, are often involved as a part of the project disciplines. It is necessary for the team to assign modelling roles like requirements specifier, business process designer etc., to the participating stakeholders. The activeness of the stakeholders is related to less requirement of feedback in the Agile Unified Process.

> Agile Modelling Standards: A significant part in AUP is played by the UML (Unified Modelling Language) diagrams. For maintaining its agility, the agile teams often blend the standards and the guidelines together. On the other side, in an RUP project, the guidelines to be adopted by the teams for creating modelling artifacts are included.

> Gentle application of the patterns: The AUP teams get the full freedom for choosing which modelling patterns to use. However, in RUP, these patterns are defined by the product depending on the modelling disciplines being followed. This practice has led to an enhancement in the performance of Agile Unified Process by easing the way to apply patterns. But, this concept is not as explicit as it should be.

> Application of the right artifacts: The advice for creation of various types of models is provided by the AUP and it is one of its strengths. The recent version of the RUP also provides plenty of advice on creating non – UML artifacts (UI flow diagrams, data models etc.).

> Collective ownership: This concept of Agile Modelling is used for making enhancements in the projects developed using AUP. But, it has to be assumed that the open communication is supported by the team culture. Along with supporting this concept, AUP lays strong stress on issues concerning configuration management. Therefore, the change management processes sometimes can be a hurdle in path of development.

> Parallel creation of several models: This is an important concept of UP. The team is required to check the activitiy diagrams corresponding to each discipline and see if they are being worked up on in parallel. But there is an issue with UP which is that the flow diagrams do not explain this well.

> Creation of the simple content: The simplicity is assumed by the developers. The team needs to adopt the guidelines stating the use of simple models and also the customers must be happy with this. However, many organizations often find it difficult to adopt this culture.

> Temporary models should be discarded: The AUP team is free to decide which models to discard and which models to keep. Travelling light also helps in maintaining simplicity.

> Public display of models: Models should be displayed publicly for facilitating open communication. This way all the artifacts will be available to all the stakeholders

Saturday, October 18, 2014

What is Agile Unified Process (AUP) ?

Rational Unified Process, when simplified, gives rise to AUP or Agile Unified Process. Its developer – Scott Ambler, describes it as a simple and very easy to understand methodology for development of application software for business. The agile unified process makes use of agile concepts and techniques but remains true to its origin i.e., the Rational Unified Process. Various agile techniques are employed by Agile Unified Process for developing software:
> Test driven development (TDD)
> Agile modelling (AM)
> Agile change management
> Database refactoring

All these techniques help AUP in delivering its 100 percent productivity. In 2011 the AUP was considered to be 1 percent of the whole agile methodology. In 2012 DAD or Disciplined Agile Delivery Framework superseded the AUP. Since then most people have stopped working on Agile Unified Process. AUP is different from RUP in the sense that it works only on 7 principles:
> Modelling: Involves understanding about how the business is organized around the software and the problem domain of the project, and then identifying a feasible solution for addressing the problem.
> Implementation: The model is transformed in to executable code and basic testing i.e., unit testing is performed on this code.
> Testing: An objective evaluation is carried out for ensuring that the artefact is of high quality. Testing is done for rooting out the defects and validating whether the system works as desired or not. It also includes verification of the requirements met.
> Deployment: The delivery plan for the system is laid out and executed so that the product reaches the customers.
> Configuration management: Managing the access to various artifacts of the system includes tracking the various versions at regular intervals of time, managing and controlling the changes in them.
> Project management: It includes directing the activities of the project to the right track. These activities include risk management, assigning tasks to people, tracking their progress, coordinating with people for ensuring that the system is delivered on time. This is also for making sure that the project is completed within budget.
> Environment: Involves providing continuous guidance i.e., standards and guidelines and tools for ensuring process is going on right track.

The agile unified process follows certain philosophies as mentioned below:
> The team knows what it is doing: People don’t find it convenience to read highly detailed documentation. Timely training and good guidance is accepted by all. An AUP provides links to a number of details if you want, but it does not force you.
> Simplicity: The documentation is kept as simple as possible without going into too much of detail.
> Agility: For maintaining agility, the AUP should conform to the principles and values mentioned in the agile manifesto and agile alliance.
> Emphasizing on high – value activities: Only the activities that affect the project are actually considered. Rest of the activities are not counted.
> Choice of tools: In Agile Unified Process any toolset can be used. However agile experts often recommend using simple tools appropriate for the project.
> The agile unified process can be tailored specific to the needs of your project.

There are two types of iterations in agile unified process as mentioned below:
> Development release iteration: For deployment of the project to demo – area or quality assurance.
> Production release iteration: For deployment of the project to production unit.

These two iterations are a result of refinement of RUP. The RUP’s modelling, requirement and analysis disciplines re encompassed by the disciplines of agile unified process. Even though modelling constitutes an important part of agile process, it is not the dominating factor. 

Facebook activity