Project Management Central

Please login or join to subscribe to this thread

Practice Areas: Agile, IT Project Management
How to estimate Bug/Defect Fix effort?for software maintenance?
Network:188



How to estimate Bug/Defect Fix effort for software maintenance ?

We do look at the past trend of the team for a period Ex: In a month or 3 Months how many team was able to fix different severity bugs/defects,Kind of velocity based on that we come with projection/forecast

Is there any other scientific method for this ?
Sort By:
Network:1092



In the same way you estimate the creation of the piece of software. There is not other way. For example, if you use funcion points you have to use it.
...
1 reply by Ruben Dario Abello Medina
Mar 09, 2017 8:08 AM
Ruben Dario Abello Medina
...
Sergio is right
And if you know the team average for resolve issues, and combine it with functions points, you can make a better stimation
Network:211



Mar 09, 2017 7:58 AM
Replying to Sergio Luis Conte
...
In the same way you estimate the creation of the piece of software. There is not other way. For example, if you use funcion points you have to use it.
Sergio is right
And if you know the team average for resolve issues, and combine it with functions points, you can make a better stimation
Network:2039



If the team has been working issues for some time, then they will have a rough estimate on how long it takes to fix an issu. If you have someone triaging issues as they come in, they can assign a point system to the issues as they are assigned. Once you have the point system established and in use for a while, a velocity will start to emerge. There are many methods to use, the most important part is to use the same rating/scale/points for the work and then you can see trends and establish velocity.
Network:60513



Liane input is right
What is your track record to resolve bug\defect
Network:188



Thanks for the valuable replies , As mentioned in the question we do estimate currently by looking into pas t trend/historical information , We do have some kind of triaging/scaling mechanism for each defect as in when they arrive.

Unlike New Development there is always a challenge to estimate defect fixes because we are not building some thing new where we can break up/decompose the activities and do estimation

Defect fixing has diagnosis steps but depends on the nature of the defect Ex: Performance , Security , Functional failure , Work-flow , Environmental , Configuration , …etc

Simple defect may take days/months to fix , Complex defect may take couple of Minute/hours to fix it, So our trend is fluctuating - + 25 % over period time,

Typical FP, UCP, Story points, T-Shirt Sizing has been tried not much help compare to past trending/Historical information

This is based on historical information I was wondering “No of defects per lines of code” or any other method/option to try out.
...
1 reply by Sergio Luis Conte
Mar 09, 2017 10:39 AM
Sergio Luis Conte
...
Sorry but I do not agree with this: "Typical FP, UCP, Story points, T-Shirt Sizing has been tried not much help compare to past trending/Historical information". In fact FP is the best way to estimate mainly software products because the amount of information inside the oficial website. And FP is used to calibrate others like UCP or Story Points. What you have to consider is that all related to maintenance must not be considered in isolation. You need a maintenance strategy and process similar to other industries. For example, what kind of maintenance will you perform? Bug fix is not a good strategy. Is about quality control. And you have to pointed out to quality assurance. Here is where you have to prevent instead of cure. And things like preventive maintenance deliver more value to organizations. But, at the end, the estimation methods are the same. At least, that is what I used and researched after more than 30 years working in the field.
Network:1092



Mar 09, 2017 10:28 AM
Replying to S Rajasekar
...
Thanks for the valuable replies , As mentioned in the question we do estimate currently by looking into pas t trend/historical information , We do have some kind of triaging/scaling mechanism for each defect as in when they arrive.

Unlike New Development there is always a challenge to estimate defect fixes because we are not building some thing new where we can break up/decompose the activities and do estimation

Defect fixing has diagnosis steps but depends on the nature of the defect Ex: Performance , Security , Functional failure , Work-flow , Environmental , Configuration , …etc

Simple defect may take days/months to fix , Complex defect may take couple of Minute/hours to fix it, So our trend is fluctuating - + 25 % over period time,

Typical FP, UCP, Story points, T-Shirt Sizing has been tried not much help compare to past trending/Historical information

This is based on historical information I was wondering “No of defects per lines of code” or any other method/option to try out.
Sorry but I do not agree with this: "Typical FP, UCP, Story points, T-Shirt Sizing has been tried not much help compare to past trending/Historical information". In fact FP is the best way to estimate mainly software products because the amount of information inside the oficial website. And FP is used to calibrate others like UCP or Story Points. What you have to consider is that all related to maintenance must not be considered in isolation. You need a maintenance strategy and process similar to other industries. For example, what kind of maintenance will you perform? Bug fix is not a good strategy. Is about quality control. And you have to pointed out to quality assurance. Here is where you have to prevent instead of cure. And things like preventive maintenance deliver more value to organizations. But, at the end, the estimation methods are the same. At least, that is what I used and researched after more than 30 years working in the field.
...
1 reply by S Rajasekar
Mar 09, 2017 10:58 AM
S Rajasekar
...
Yes , I agree with your point we need to focus on Defect Prevention and Quality Assurance

We calculate defect escape rate and root cause analysis based on that we take action items to Development , Quality…etc team

1.Strength Unit test , Improve system/domain knowledge for engineers , Re-factoring ,Peer Review..etc
2.Test case coverage, Error, Improve system/domain knowledge..etc
3.Better documentation , Training to support team, Improve trouble shooting skills ..etc

These are continues process at the same to time we also need to fix incoming customer defects and respond to market with Service pack and hot fixes..etc

Right Now we fix defects and do quality checking and release fixed defects solutions to market in regular interval (Ex: Monthly/ Every Fortnight ), In fixed interval what is done we release them

We would like to move to a model where we can estimate list of defects to be fixed with estimate and come up with delivery dates. We want to estimate/decide items that can be delivered in fixed interval .
Network:188



Mar 09, 2017 10:39 AM
Replying to Sergio Luis Conte
...
Sorry but I do not agree with this: "Typical FP, UCP, Story points, T-Shirt Sizing has been tried not much help compare to past trending/Historical information". In fact FP is the best way to estimate mainly software products because the amount of information inside the oficial website. And FP is used to calibrate others like UCP or Story Points. What you have to consider is that all related to maintenance must not be considered in isolation. You need a maintenance strategy and process similar to other industries. For example, what kind of maintenance will you perform? Bug fix is not a good strategy. Is about quality control. And you have to pointed out to quality assurance. Here is where you have to prevent instead of cure. And things like preventive maintenance deliver more value to organizations. But, at the end, the estimation methods are the same. At least, that is what I used and researched after more than 30 years working in the field.
Yes , I agree with your point we need to focus on Defect Prevention and Quality Assurance

We calculate defect escape rate and root cause analysis based on that we take action items to Development , Quality…etc team

1.Strength Unit test , Improve system/domain knowledge for engineers , Re-factoring ,Peer Review..etc
2.Test case coverage, Error, Improve system/domain knowledge..etc
3.Better documentation , Training to support team, Improve trouble shooting skills ..etc

These are continues process at the same to time we also need to fix incoming customer defects and respond to market with Service pack and hot fixes..etc

Right Now we fix defects and do quality checking and release fixed defects solutions to market in regular interval (Ex: Monthly/ Every Fortnight ), In fixed interval what is done we release them

We would like to move to a model where we can estimate list of defects to be fixed with estimate and come up with delivery dates. We want to estimate/decide items that can be delivered in fixed interval .

Please login or join to reply

Content ID:
ADVERTISEMENTS

"Laughter is the shortest distance between two people."

- Victor Borge

ADVERTISEMENT

Sponsors