Fake Email Server for TDD

I found this nice tool called Dumbster which is a fake SMTP server. This can be used for unit and system testing applications that send email messages.

Agile 2.0 ?

Recently buzz about Agile 2.0 is going around in agile community, looks like sponsored by Microsoft. Here are my thoughts about Agile 2.0

1. From the above article, it looks like MSF is concentrating on agile practices. This would the right recipe for failure. Agile manifesto never talked about practices but only on principles and values which are global and constant. Practices needs to be tailored to individual project environments based on principles and values. A practice that works in one environment may not work on the other

2. Name Agile 2.0 itself is a misnormer. In continuation of point 1 above, there was no such thing called Agile 1.0 before !! We had just agile principles and values.

Law of diminishing marginal returns

I found this nice article on wikipedia about law of diminishing returns.

Quoting from wikipedia:

The "law" of diminishing marginal returns says that after a possible initial increase in marginal returns, the MPP(Marginal Physical Product) of an input will fall as the total amount of the input rises (holding all other inputs constant)

Here are some of the examples from software development environment where we can see the above law in practice:

1. Addition of new team members to the project "seems" to increase productivity marginally but reduces over a period of time.
2. Developers excited about implementing the new technology, and after a while losing interest in it. For ex: implementing Webservices

Software development and mass production mentality

In the last few decades the mass manufacturing industries went on a spree to create "intelligent" machines. The intention of this creation was to solve complex industrial problems quickly, and reduce the manufacturing defects. The only thing the "isolated" worker would do in the assembly line, was to pickup the parts from one line and feed it to next. There was no need to "talk" or "communicate" to the fellow worker. Can you see any similarity between the above example and, "traditional" software development teams ? . The traditional software development is plagued by "mass production mentality". The requirements team would feed set of documents to analysts, who in turn analyze and feed to design team. The design team would feed the high level and low level design docs to development team, etc. Each team has there own specialty and is happy doing the routine work !.

In the lean enterprise, the employees would talk and take help from each other while solving complex problems. They would make use of big visible "andan" systems to constantly monitor the status of production, defects, inventory information. This leads to better communication, leading to creative solution.

I have become Scrum Master HEY !!!

Craig conducted the course on "Scrum Master Certification" in Taj Gateway on Residency road. We were around 40 people, mostly from Valtech India. It was fun and also learnt a lot. Since, I have been working with Craig from the last 6 months, I knew many of the concepts and much more. I was really moved when Craig gifted me the Critical Chain book. Looks like he is asking me to start looking towards next wave coming up in process improvement !!

Finally I can call myself as "SCRUM MASTER" HEY....HEY.. I am just excited about this.

Agile journal article

My article on Benefits Of Tool Integration In Distributed Agile Development was published, and so far received good feedback from friends and well wishers !

Technology experts rescuing process team

From the last couple of months, I have been working closely with Craig Larman and I got a chance to learn not only about Agile methods but also a lot about OOA and OOP. I am also a big fan of people like Martin Fowler, Ron Jefferies, Jeff Sutherland who have contributed a lot to Agile methods. If you carefully observe their past life, they all come from technological background rather than the pure process background. This is leading me to believe that a person with both technology and process background could contribute more to software development world than being purely in one of them. I would like to quote one of the real world examples, I was having a chat with group of SQA team, and still they think TDD is testing rather than a design !!. I saw today Craig checking the code of a team member and smoking code smells out, and I don't think this could be handled by a "pure" QA person.

Motivation behind scrum meetings

Scrum meeting is just not "Standing up" ,"meet" and ask 3 questions. There is certain motivation behind this meet. While reading an article I found one of the motivation to be:

When one team member shares an obstacle in the Scrum meeting, the entire team’s resources come together to bear on that problem. Because the team is working together toward a shared goal, every team member must cooperate to reach that goal. The entire team immediately owns any one individual’s problems. [Rising and Janoff, 2000]

This is also in accordance with the lean thinking, Toyota way !!

Agile Pit fall

Even though Agile methods are supposed to reduce the risk of the project, improve the quality of the deliverables, one thing I feel which would complement agile is "measuring the quality in each step". Agile principles and values never talk about any particular "measurement", and the Agile methods who have tons of practices have not given much importance to "measurable result".

Quality

I am sure you would be using the word "quality" hundreds of times a day if you are in software industry. Even though there are various definitions of quality from various school of thoughts, for example:

Quality is conformance to requirements.

and the other one here

Quality is the capacity to remain robust and conformant, while adapting to new requirements.

.

I strongly agree with the second definition. The reason being, I can write my program to conform to requirements but what if, the design is poor ? what if the code is crappy ?

Toyota Way and the Stand up meeting

I remember somebody telling me that, in Toyota manufacturing plant and in the assembly line if somebody discovers some fault in the product, immediately they would raise an alarm. This in turn signals everybody to stop what they have been doing, and approach the person who raised the alarm. All of them in that shop floor in turn would discuss and help the person in resolving the fault he/she has seen.

I believe the same principle is being applied in our daily stand up meeting. Most of the time, the daily stand up meeting is considered as "status meeting". In fact it is not, especially if you look at the third question "Any obstacles stopping the developer from proceeding further?". The reason I believe behind this question is to make the other team members aware of this issue, and in turn all of them can jump together in helping this person out. This is in line with the Toyota manufacturing analogy shown above.

Until we don't understand the principles behind practices, practices remains as practice !!

Discussion with Craig Larman

Nowadays I am working closely with Craig to get a deeper understanding of Agile and Scrum. Here is the summary of key what I understood

1. There is nothing called agile "method". Agile provides principles and values. There are agile methods like XP,Scrum,etc

2. Most of the time people who go through scrum master certification end up becoming "Project managers having Scrum master certification"

3. Scrum Masters are obstacle removers in a project rather than managers controlling the project. I can clearly see in my surroundings, how difficult it is to achieve this particular goal. The reason being most of the people are coming from "Project Management" background !!

Becoming a scrum master is not "practicing some things", it needs total change of mindset, the way we lead life at work, the way we think.

I have started working on implementing "Virtual Lava Lamp", and it looks like a lot of configuration changes needs to be done while integrating Cruise control with lava lamp.

Good quote from Ron Jefferies

The greatest mistake we make is living in constant fear that we will make one.
-- John Maxwell

Why some teams hesitate to maintain charts

I was reading Jim Shores blog, and came across an article where he has recommended to put some thought process if your team hesitates to update the "Big Visible Charts". I am not going to rephrase what he already said, but going to paste his words here:

he first question to ask is, "Did the team really agree to this chart?" An informative workspace is for the team's benefit, so if team members aren't keeping a chart up to date, they may not think it's beneficial. It's possible that the team is passively-aggressively ignoring the chart rather than telling you that they don't want it.

If people won't take responsibility, perhaps you're being too controlling.

I find that when no one updates the charts, it's because I'm being too controlling about them. Dialing back the amount of involvement I have with the charts is often enough to get the team to step up to the plate. Sometimes that means putting up with not-quite perfect charts or sloppy handwriting. I like pretty charts, so that's hard for me to do... but it pays off.

Manual Testing Vs Automated Testing

I have seen posts on the web debating about ONLY in favor of automated testing. My view is we need to have a balance between the automated and manual testing. Couple of points

1. In order to have the tests automated and working with cruise control, a sort of maturity and experience is needed. I have tried implementing TDD with my team, it took really a long time for them to grasp.

2. Even if somebody is trying to automate testing, till date I have not seen anybody implementing them for the entire features. What I have observed is, developers would start writing unit tests and at some point of time, either to meet deadlines or for some reason they would skip writing tests and get into coding straight.

I always think it is a good idea to automate as many tests as possible and also keep manual testing continued.

Manual Testing Vs Automated Testing

I have seen posts on the web debating about ONLY in favor of automated testing. My view is we need to have a balance between the automated and manual testing. Couple of points

1. In order to have the tests automated and working with cruise control, a sort of maturity and experience is needed. I have tried implementing TDD with my team, it took really a long time for them to grasp.

2. Even if somebody is trying to automate testing, till date I have not seen anybody implemeting them for the entire features. What I have observed is, developers would start writing unit tests and at some point of time, either to meet deadlines or for some reason they would skip writing tests and get into coding straight.

I always think it is a good idea to automate as many tests as possible and also keep manual testing continued.

EJDC 2006 Conference

I was planning to post this blog from some time.
on 25th I had been to EJDC conference as a speaker and shared my thoughts on Agile best practices and metrics at EJDC Conference. It was supported by Kushal a non profit organization helping poor children in India.

My presentation was mostly towards the best agile practices that are in town and also some of the most popular agile metrics. I got my presentation reviewed by Craig Larman on thursday and got very good suggestions on usage of some of the "terminologies".

Overall I feel the presentation went very well. First half of the conference was devoted to technology related topics(mostly on SOA).

Humane Interface Vs Minimal Interface

I am very new to Ruby(In fact I am trying to understand what it is), and today I stumbled upon an article by ' Martin Fowler ' which is an eye opener for me and I beleive, this could be a good start for me to start visualizing why there is so much hype about Ruby !! I have suddenly started seeing coding in Java/J2EE as x86 programming !!

Traditional "Done" Vs Agile "Done"

I found this interesting post on the web. Here the author williamcaputo is trying to describe the "thinking" of "done" by the traditional team Vs Agile team.


Traditionally 'done' means "We won't have to work on that anymore - and if we do, somebody must've screwed up." For Agilists 'done' means "We have accounted for what is known. If the situations changes, we expect to revisit this work."

Thus the two approaches have different goals: One is to finish, the other is to keep pace with change.

How Idiotic is to track % completion of the project

In recent days, I have been working closely with Craig larman to bring out a new domain modal based on scrum for Valtech cockpit. During the discussion with Craig, he talked about how stakeholders ask for % completion of the project and, how our PMs give them a %(percentage) !! It all looks idiotic if you read the rest of the following section.

The core reason for introducing agile methods in software is to manage the changes, adapt accordingly and be open to get changes. As per the statistics, change requests increases as the project size increases. If we are getting more changes obviously it would add more effort to existing estimated effort and this becomes a moving target. How can anybody estimate how much % of project has completed in this scenario? It is as good as telling not the truth, if I tell my stakeholder that we have finished 50% project !! (Unless you add a clause saying "as of today")

So, instead of using "% Complete" as the measurement, one can use the "estimated effort" remaining to complete the project