Saturday, November 21, 2015

Is Scrum (and probably Agile) Dead?

Recently, I came across an interesting article: Scrum is dead: breaking down the new open development method. I usually do have big problem with such big declarative statements like 'Scrum is dead'. Actually, it's pretty easy to refute the whole rationale by bringing about one counter example!

But, to be frank, the article is good, with the author trying to put some basic principles for the profession of software development; principles that I fully agree with and promote. On the other hand, I strongly disagree with some very fundamental ideas in the article:

Agile is Not a Methodology


First, I find the author builds upon a fundamental idea that "Agile is a methodology". Well, this is incorrect. Also, the article conveys an idea that Agile and Scrum are two sides of the same coin (a misconception about Agile. May be the author doesn't mean that, but this idea is there in the article). Actually, Agile does not prescribe any methods. Rather, there are methods which are built while Agile in mind. So, what is agile? I always preach Ahmed Sidky's definition of Agile:

"Agile is a mindset, that is in the software world, established through 4 values, grounded by 12 principles, and manifested through many many different practices"

In this sense, there is no point of discussing whether Scrum, Kanban, or Open Development are useful methodologies or not. Because the whole idea of usefulness is very situational. What is useful in one situation is not useful in another, what makes sense for one team is totally irrelevant for another. 


Is Open Development an "Evolution of Scrum"?


The other statement which I find misleading is that Open Development is an "evolution of Scrum". Well, this is also incorrect. The Open Development description as appears in the article is totally in accordance with the Agile mainstream since its inception. The stream upon which ideas like Scrum and XP was built. In such mainstream, major advances in Software development were invented, such as automated tests and Continuous Integration and Deployment; and major ideas were revived or re-established like IID (Iterative and Incremental Development) or self-organizing and cross-functional teams.

Well, when to use which? or shall we use None?


Having said all of that, I do admit that Scrum (as per the official Scrum guide) is not suitable in many cases. A better statement is that: Iterative development is not useful in many cases. Some project types cannot be led using timeboxed iterations. Usually, the reason is that the demands of the business and the development pace are much faster than a one-week timebok, which is the smallest iteration timebox. The idea is very simple: If business needs something now, and development is capable of providing this thing at this now, why wait till end of iteration, which may be 2-3 or even more than a week ahead?!

On the other hand, I recently came to a situation where I couldn't use Kanban to manage my project. The unknowns of business are so much, and the availability of the business people is limited; besides great technical risks which needs prolonged time of team collaboration and deep thinking. In this specific case, we switched from Kanban to Scrum.

Managing Flow or Managing Iterations?
Actually, what we have done is to move not from Kanban to Scrum, but from Managing Flow to Managing Iterations. Kanban and Scrum nowadays are metaphors for two more fundamental management ideas: Managing flow versus managing iterations.

So, In this project instance, I found that putting constraints on flow is stopping us rather that helping us. I found that a timebox of one or two weeks gives the team more room to act, react, and innovate. In the meanwhile, it gives us some very high level indicators about team progress.

When shall we use None?
One final thought is that even if you're doing Scrum or Kanban, it's just a starter. It just brings about a sense of discipline and cadence which enables the team to measure it's progress and improvement over time.

So, if this is only a starter, you should expect that teams will move to other practices and probably invent their own process in the near future. At this time, I would say that this team is really Agile!

Monday, October 26, 2015

Things are Inverted in Agile!




It seems that many things in Agile are INVERTED !!

I noticed two important aspects (so far):

Project Startup Activities

We used to code then deploy. Just before deployment, we figure out how are we going to deploy, on which server, testing environment or staging environment. Usually, we may wait for sometime till these logistics are sorted out. 
In agile, things are different. We start by preparing the deployment environment, set it up, link it to our code deployment scripts using some Continuous Integration and Deployment tool. 

The Famous Design Phase

"In TDD, there is no design stage; design is inverted and distributed.", I quoted this from the online TDD training from Industrial Logic. It is fairly true and powerful concept. 
The idea is to code-test-design (or refactor) instead of design-code-test. Well, there has to be some upfront critical architectural decisions, but these are very few and takes several hours to several days at most.

Saturday, September 26, 2015

Dead Code; Is it Really Harmful?

The answer is YES, and can put you out of business in 45 minutes, like what happened with Knight Capital Group!

Recently, I have been watching uncle Bob explaining his famous SOLID principles, and he mentioned a very interesting case of a bug. The code mistakenly set a flag which enabled the execution of a piece of dead code.

The side effect of this execution is 440 million USD's losses which is 5 times the market value of the company :)

Further read about this case:

http://www.bloomberg.com/bw/articles/2012-08-02/knight-shows-how-to-lose-440-million-in-30-minutes

Tuesday, July 14, 2015

Lean CMMI - My Presentation at Agile2015!


First and foremost, I'm so grateful and thankful to Allah (swt) for helping me getting through this research paper. Not only that, I got it accepted in the Agile Conference 2015; the biggest Agile conference in the world.

I have published several papers before, but they were all industrial reports. Industrial reports (aka experience reports) document useful experiments done in the industry. As opposed to research papers, industrial reports do not necessarily use a scientific method to prove something.

My personal opinion, Industrial Reports are very useful, and probably more valuable than research paper. They reflect real experience and one can draw benefit from it right away. So, if your audience are software practitioners, then research papers are not the right means to communicate your experience to them :)

If this is the case, why did I write a paper? Two reasons:
  1. The first reason is that I had data! and found it interesting to experiment writing a paper.
  2. The second reason is to write a proof that incremental approaches in process improvement is better that holistic approaches. 
Here is a link to the abstract of the paper: Lean CMMI: An Iterative and Incremental Approach to CMMI-Based Process Improvement.

In this paper, I'm comparing results of process improvement in Configuration Management process area of CMMI. The first group of results are based on traditional process improvement effort (one process area at a time). The second group of results are based on the 'Process Increments' method, an iterative and incremental approach for software process improvement. This model divides improvements into small chunks as in the following model:



These Process Increments are dynamic. You can add, remove, re-prioritize them as needed. It's like user stories in software projects.

The results of the study are in the following diagram. It's a box plot of SP ratings in semi-formal CMMI appraisals. As you see, the ratings in the second group of companies are much higher than ratings in the first group of companies. is as follows:


Those who knows six sigma or has experience in research would know that these results are real proof that iterative and incremental methods works better!

Saturday, April 25, 2015

8 Months of Being Self-Employed!

Today is my 8th month after resignation from SECC and becoming self-employed at Agile Academy. This was a challenge full of new experiences, problems, obstacles, failures and successes :)

It's like any new startup experience, in which you're doing everything which you used to get them done by whoever else! Many things changed in my life like being with my wife and children for a longer time, setting in front of my laptop for a longer time, thinking about business and opening new opportunities. It's an experience in itself, not to be expecting any income at the beginning of every calendar month .. :(

One thing I noticed is that what Allah gives you cannot be compared to your effort. Of course you have to exert all what you can so that Allah give you, but it's always Allah's will, and I felt so relieved when I finally understood that Allah will not let you down as long as you are doing all what you can .. Al7amdulillah.

One very important thing to consider is that you have to always do what you like and what you know. Both are important. Then, الرزق will come along the way, whether through direct work you produce or through any other route.

One more thing .. It's not appropriate to work day and night. You have other obligations in life which you have to fulfill, like your parents, your wife and children, your own self, and before all of that, your Lord Allah SWT.

Friday, March 13, 2015

My Article about Technical Leadership

My second article published at infoq is for Technical Team Leaders. It's a post which I filled for many many years, and I have collected all my observations about what should a TTL be doing with the team in this article.

http://www.infoq.com/articles/technical-leadership-overseen

good read :)