Wednesday, October 11, 2006

Balence, Tranquility, and SOAP

I am behind schedule with my work. I attribute a good bit of this to my vices as a developer, and just as much of the problem to my good attributes. I place an equal portion of the blame on SOAP.

Striving for the goal of Beautiful Code, we can find ourselves lost on the way to actually writing something that gets the job done. Throwing away perfectly working code, because an alternative way to achieve the same results is more elegant isn't something that we might see as a bad idea. If the code is more l33t now, it will give us less trouble tomorrow when we need to port it to my toaster. We'll use anything to justify the overworking for code beauty.

Is it always worth it? How beautiful is enough and when are we just wasting our time (and money)?

There are terms thrown around like "elegant" and "pythonic" to measure the quality of code with no attention to the code actually reaching the goal it sets out to perform. The code may work, but that doesn't make the code good code. Without a sense of time, scale, and the big picture, the search for good code can overshadow any good developer's work towards working code.

However, as any issue as a flip side, those developers getting lost are doing so in the name of a good fight: the first against bad code. We might get lost and never complete our code, or complete it late, but we do so with the complete belief that it was worth it. The code took several weeks longer to develop, but just look at how beautiful it is. Without the struggle for good code, our working, bad code would eventually overshadow us just as much and consume our time with maintenance, refactoring, and the mother of all frustrations in coding: trying to read your own work.

I am wrapping up some SOAP-heavy work and the path to completing it has been a testimony to the struggle of balance in code. Recent refactorings of the actual SOAP response processing ended with a good chunk of bad code. I don't like the way I'm doing lots of things, or the fact that it doesn't parse corner cases the service I'm using doesn't even use. The code is not the beautiful code I would like to call my own, but the code is working code and does everything it needs to do. I had to bite my own hand to keep the refactoring to a minimal and focus solely on the aspects of functional goals, ignoring aesthetics.

Be careful on the road to good code. Somewhere along the way, you can easily get lost and never reach the point of having actual, working product. Sure, the code will be incomplete, but it will be a fragment of beauty. Learn the value of a completed mediocre code set over the eternal development of more beautiful code, which does exactly the same thing.

1 comment:

Anonymous said...

HEAR HEAR!

Far be it from me to defend PHP, but those guys actually probably get more done then I do because they're willing to live with "bad" (ugly, inelegant, possibly even insecure or slow) code that still works!

This is indeed a balancing act... one that, if we are to be successful geniuses, we must learn and learn well. normally we can throw this back over the fence toward the project manager... but when the project manager is us, we're our own worst enemies!

isn't hacking fun?!

-j

I write here about programming, how to program better, things I think are neat and are related to programming. I might write other things at my personal website.

I am happily employed by the excellent Caktus Group, located in beautiful and friendly Carrboro, NC, where I work with Python, Django, and Javascript.

Blog Archive