Quality in software development

What does this truly mean? Somebody actually told me, during a business talk, that quality was just a buzzword. It felt wrong, but in a way it felt real too. I now understand where he was coming from and I can explain it. The only way he experienced quality in software development, was through the word itself: quality. The context around it was usually messy (code, defects, unhappy clients etc.). No wonder he started hating the word and sarcastically referring to it as a buzzword.

So back to the question: what does quality in software development really mean? There is more than one point of view that we should explore, so here we go:

The big picture

Ultimately, everyone wants to get what they ask for, without any hassle and within their budget. The holy grail is getting the best thing, for free, right away. Don’t try to fight this. It’s true. Think of the best version of anything you want and tell me you wouldn’t like it materialising in front of you, right now. So from this point of view, quantification of quality is done on the scale of “this is not what I wanted and it’s too expensive and it takes too long to get” to “this is exactly what I wanted and it’s free and I got it“.

Of course, there are some variations, one of which is heavily encountered in the industry: “this is not quite what I wanted, but it’s cheap, yet it took too long to get“. These are the murky waters many clients jump into when having to do trade-offs. In my experience, these are usually budget trade-offs, that end in an unfortunate state, where over-budget gets spent in the same bucket, but for patch work.

This is the scenario I like best:

Client: This is my problem. Can you solve it?

Provider: Yes.

Client: What will it cost me, in time and money?

Provider: I estimate around 6 months and 500.000$

Client: Great! Let’s do it!

(every couple of weeks)

Provider: Is this what you had in mind?

Client: Almost. Here, I’d like a little more blue.

Provider: Done!

(around 6 months later)

Client: It looks like we’re done. Great job! Thanks!

Provider: Anytime.

There are lots of details behind this scenario, but this is the big picture. When I read it, it gives me a peaceful feeling of “yes, this is how business should be“.

Processes

These are blueprints of behaviour that help when dealing with certain problem patterns. They also help the entities that participate, have some (illusion of) predictability of the future. It’s a sort of guarantee that the future will happen in a certain way. This helps people feel in control and reduces the level of stress. BUT! Since we derive a level of comfort from having the illusion of a certain future, we tend to equate the process with the future itself. This is a mistake that leads to process rigidity and reluctance to step outside its bounds.

I like to look at processes as checkpoints and guidelines along the road. The main focus should always be the product. Engaging my expert skills right now, while actively building it, IS the best guarantee I can have of the future I want. The future is built from a constant stream of present 🙂

I always look at the reason for the existence of the process, so it gives me a better understanding of why it was set up in the first place. This way I can work along the lines of that reason, rather than blindly following a process I cannot understand (and sometimes come to hate).

One thing I noticed is that the fuzzier the goal, the more processes there are. The more processes there are, the more people tend to make a goal out of them. And so the wheel spins…

Craft

This is, surprisingly, not the most looked at factor in the software development industry. In my experience, if a client sees processes and pays an amount that suits her, THE CODE (which is ultimately the product) doesn’t matter as much, as long as it… exists.

WHAT??!! This is outrageous! But wait… is this something the client should even be aware of? Why would it be so? If I buy a pair of shoes, I’m not interested in the way they are built, what I’m really interested in is how they look, how they feel and how long they last.

The craft IS really important, inside the guild. Because we’re talking about code, things should be simple, really: it should work as required (without maintenance, preferably) and it should be easy to modify (if needed).

These are two seemingly straightforward characteristics, but to achieve them, takes not only time and experience, but continuous learning too. A craftsman should be able to easily understand the request (even help with it) and pick the most appropriate tools for the job. A craftsman should be able to read code to business people so that they hear a narrative. A craftsman should be able to write such code.

I want my client to understand where their product stands at any time and what it can do. I do NOT want my client to be forced to listen to what kind of refactoring and how many unit tests I wrote yesterday. Or even worse, she shouldn’t be forced to decide if I need to write them tomorrow.

Also, I don’t want to hear that “we are thinkers, don’t force us to estimate“. Well, if you’re thinkers, then think about how long it takes you to do something you should already know how to do. Nobody said estimates are deadlines (did you?!!), but clients need to have a rough idea about resources they need to allocate. Your GPS doesn’t say “you know what… we’ll get there when we get there! The important thing is the magic I constantly use to display these awesome maps.

Trust

I don’t want to be cheesy and say that trust is earned, because of course it is. The difficult thing is to trust someone I don’t know. Looking at references and traces of past work helps, yes, but I also place a huge weight on the quality of the first discussion. It always served me well in assessing my future collaboration with anyone. Everyone is an intuitive psychologist by nature (not mentioning special health conditions).

The thing is, we almost always know (trained people are an exception, hence the “almost”) when someone is hiding important details in a conversation and we can also “feel” when what they say is what they think.

Obviously, the higher the level of trust, the higher the expected product quality. I’ll stop here and save the psychology magic for another time 🙂

Conclusion

These are my most valuable points of view, when looking at quality in software development. I also look for all of the above in business partners and also, I’m happy to say that for us, at bitGloss, quality is not just a buzzword.