Going Into Design Debt

0 comments suggest edit

Found this post in the trackbacks section of my post on Mental Laziness. It’s a classic example of where management pressure often leads to mental laziness.

In this scenario (go read it, it’s short), the ViewState of a system they were working on was extremely large and causing problems for the client. The quick solution was to perhaps use some sort of Http Compression or ViewState compression.

But the author of the post, David, suggests to his coworker that they should dig into why the ViewState is so large as the ideal course of action.

In this situation, with management breathing down your neck, it is prudent to spend some time digging into the root cause of an issue. Often, you’ll end up finding an obvious mistake and dramatically improve the application. But at the same time, you want to be careful not to spend too much time banging your head against a problem when you have a workable (albeit band-aid) solution in your hand.

In their situation, it makes sense to set a time limit to investigate the root cause. Perhaps the root cause is that the pages are just plain big due to requirements, and there is no “mistake” to correct. In that situation, the right solution IS ViewState compression. The extra investigation time bought you that assurance.

Suppose they couldn’t find the solution in that time span. At this point, there is nothing wrong with just putting the compression solution in place without having determined the root cause. However, doing so will incur a Design Debt, and that has to be considered when making the decision.

In the book Refactoring to Patterns, Joshua Kerievsky talks about the concept of Design Debt. Design Debt is the state your code is put in when you write crank out code without regard to its design just to meet a deadline. There are situations when this is necessary. Sometimes you just have to bite the bullet and weigh business needs against design purity and just spew code.

As an illustration of design debt, consider building a house of cards three stories tall. You set up the foundation for three stories, you start to build it, but halfway through, your boss comes in and tells you to build it seven stories tall, and get it done ASAP.

Sure, you can do it without redoing the foundation, but to add even more stories later, you’re going to have to revisit the foundation or the whole thing will come tumbling down. Or at least, it’ll be very slow to build that next floor.

Software is much like this. If you are rushing code to meet a deadline, you will go into design debt and that debt has to be paid off, otherwise future development will often slow dramatically and be much more error prone. The obvious problem is how do you sell this argument to your boss? From the business people’s perspective, you’ve delivered working software with X number of features in a short period of time. Why can’t you deliver the next X number of features in the same period of time?

Ah. Therein lies the rub! The business people don’t understand software, and by meeting their insane deadline, you’ve created the impression that…well…that is a normal project pace.

This is where Kerievsky believes the Design Debt metaphor may be of assistance. By phrasing it into financial terms the business folk can understand, the hope is they’ll trust you and provide time to pay off the debt. It’s a simple formula, If you go into debt. The debt must be paid.

Sure, I can meet this insane deadline. But we will go into debt, and it will have to be paid before we add any more features.

Try to make that an expectation up front if you can. It won’t always work, but if it never works, consider moving to a less dysfunctional work environment.

[Listening to: Suite No. 6 In D Major, Gavotte I - II - Yo-Yo Ma - The Cello Suites - Inspired By Bach CD2 (3:42)]

Found a typo or error? Suggest an edit! If accepted, your contribution is listed automatically here.



4 responses

  1. Avatar for public MattBerther : ISerializ
    public MattBerther : ISerializ October 3rd, 2005

    Its been a few years since I've been introduced to the concept of Technical Debt. There's been some attention to this idea in the blogsphere lately. In a nutshell, when you put off doing something in your development effort, you...

  2. Avatar for Jack's WebLog
    Jack's WebLog October 17th, 2005

    One of the biggest problems I experience at work is

    design debt.

  3. Avatar for you've been HAACKED
    you've been HAACKED October 1st, 2006

    Is Backward Compatibility Holding Microsoft Back

  4. Avatar for schlub
    schlub January 21st, 2008

    Nice article - it sums up a lof of situations at my current work... I don't want to be here when the house of cards is going to fall!
    Note: I was about to correct your misspelling of 'stories' (instead of 'storeys', as 'stories' are read to small children) but I see that you're American so I'll let it slide :P