I often come across people at management positions who clearly want to manage by tracking only numbers but fail to understand why were those metrics in place. We were part of a six team project, we were asked to go through our requirements and give an estimate including any proof of concepts and study that we had to do. The estimates were in points in fibonacci but were tightly mapped to number of days. When we questioned about why points to days the answer was to keep the yardstick of measurement same across all the teams in the org. It was setup for failure, but our team had to move on with the development so we went ahead agreeing to the terms.

After the first sprint our team had missed meeting the estimate while all the other teams had achieved it, we were met by one of the senior managers who gave a stern warning to the team that points are non negotiable and wants us not to fail again. When the time for demo came by, we were the only team that did a demo of the working software and product owners were able to immediately grasp what was going on and gave feedback on it. All the other teams had met their estimated points but they did things like ‘Study ABC tool’, ‘Setup CI machine’, ‘Setup Dev machine’ and so on.

Our team did all those and also demoed a working software, the other teams have bloated the estimate and bought time. This went on for some iterations, all the other teams were able to meet the estimates but other than our team all the other teams were not able to showcase their working software. The management did not care much about working software, their success was measured by meeting the points; it was the product owners who suffered the most as it took an extremely long time for them to get something working.

When you choose to measure something people will optimise for the metrics. Velocity and points are a way to help plan and size the software so that people can be allocated and releases can be planned, when it turned out to be a yardstick then there is only movement but no progress. People will eventually game the system and it becomes a toxic cycle.

Please prioritise working software over points

During our college days (2001) when Bluetooth was in its early stages the Bluetooth SIG (Special interest group) tied up with IEEE and announced a Bluetooth based theme for CSIDC (Computer science international design competition) with good perks like Bluetooth adapters, Windows XP and Visual studio licenses for participants and good prizes. One of our friends read about this and three of us put up a proposal with our ideas and we got through the initial round. We had a lot of ideas with us to implement a prototype, it is at this point a software company which was run by one of the alumni came over to help us in planning.

Our initial plan was to implement quickly and try it out on the users for feedback.
This was for a generic identification device that can be configured to use for payments, tolls for cars, parking etc. The idea was that Bluetooth devices can hold a lot of data and gives the users the flexibility to create as many profiles as they want.

One of the seniors in the company invited us to their place, in our first meeting itself we were stumped when we saw the mountain from which the waterfall process flowed. We were slapped with a software requirements specification document template which would have taken a few weeks to fill for the amount of ideas we had. It also had various parameters to rate the requirements like return on investment etc. I could never forget the feeling that we had on that day, we dropped our idea altogether and went on to find something which had requirements perfectly documented.

By the time we finished the first cut of the project and submitted we were only half happy that we did finish something. We ended up converting EEG wirelessly and transmit through Bluetooth which was not a novel idea at all. The Waterfall process killed the idea, the fire and above all nothing useful came out of a defined process. Someone else who had submitted a similar generic id device entered the top 10 and ended up winning something. Some events leave a lasting impression on your life, this one was so strong that I had never accepted to work in waterfall development ever even though that was the only one very active around the time I graduated.

Any process followed for the sake of following it will surely kill a lot of things

I often get into a debate with people about the type and the extent of testing that has to be done for software. One of the key arguments I hear is that developers should focus on creating only the happy path and leave the rest to the testers to find or have SDETs write extensive automated tests while developers can skip unit tests and focus only on writing useful code.

I have used the swiss cheese model to explain to people about the need for layered coverage instead of concentrating all the effort onto one layer. I first heard this term while watching an air crash investigation episode in nat geo. The narrator explains that though there are millions of parts in commercial airliners, the chance of getting into an accident is far lesser than motorcycles.

How an aircraft with millions of parts is able to get back on to the sky within a few hours after landing from a long trip? The answer lies in layered testing. The pilot has a pre-flight checklist, the ground staff do a routine check for visible damages, the maintenance crew notice the hours of usage and do preventive maintenance, there is a strong network of weather monitoring and routing, ATCs manned with trained staff and modern radars, aircrafts fitted with collision avoidance systems, a long series of protocols for ground and sky movements etc.

It is the same across many engineering and medical disciplines. The risk of something slipping through all the layers of test is very rare and if it happens then it is immediately plugged. In software engineering it is a combination of static checks, unit tests, integration tests, automated ui tests, heuristics & exploratory tests, performance tests etc. No one type of test can be removed altogether and replaced by another. We need the layers to move fast and yet deliver what is required.

Image courtesy: Davidmack, Swiss cheese model of accident causation, CC BY-SA 3.0