Sacrificing time: Elegance vs crude coding

Grumpy the "made" mafia lieutenant · 2015-07-10T14:57:06

Reading the title- i know what you're thinking- there is no contest:- its Elegance coding. But wait for it.... Recently I was coding what i would classify roughly as image processing, And i had 3 marker points, roughly at the base of the image and several random lines and points. I had to scan and catch all the marker points first before proceeding to do other things (ultimate dev: bases for playing a puzzle game) with the random points and lines. Condition: Though the markers-pixels would always be roughly at the bottom/sides of the image, their positions would always vary between images, so hard-coding would never work So, the left most marker, with some trickery here and there - was fairly easy scan with rectangular shape scan. Likewise the middle-positioned marker. Both less than an hour to code. And i can describe both coding as very elegant, robust and scalable. In other words - it works independent of the marker size (number of pixels that make up the markers), scale of the image and changes in relative positions of the marker pixels. Proceeding to the right-most marker i intended maintaining this robustness. But there were problems- using the rectangular shaped scan (from the bottom-right this time) as previous wouldn't work because some of the random points would caught. Tried as much as I could there was no way with a rectangular shaped scan. Then It occurred to me that a circular shape would work - and that was the perfect solution. Now to code it. 1. I quickly used the equation of circle to generate 1/4 of a circle to scan from the bottom-right to catch the first pixel of the marker points. 2. Next i had to continue scanning, catching the pixels and as in the previous 2 markers after catching the last pixel, the algorithm should know and stop scanning, otherwise it goes on to catch the random points and messing up everything. My first algorithm didn't work. I analysed recoded, - it didn't work again and again. So i thought i needed track the way pixels are picked when the equation of a circle is embedded in a double loop. But it ended up being too bloody brain twisting and complicated Days began to tickle by, i was still struggling (was supposed to finish in just hours). It then occurred to me that there is a dirty way of doing this thing, it won't be robust, there would be some assumptions about scale , size of marker and its rough relative position. But at least i would save a lot of time (as it seems i was going no where with my coding) The delima was then do i sacrifice elegant coding and gain time by doing it the dirty way? I pondered. You know what? My EGO got in the way, i couldn't allow this thing to defeat me , so i continued seeking the solution the elegant way, Yes I eventually solved it the robust way but it took almost 4 f*******g weeks, Almost 4 weeks!!! (ok my day job got in the way a bit) But maybe thats the reason i should i have done it the dirty way- afterwards its only a hobby-baby project. Am i an egoistic person? Am i evil?? Am i unique in being like this?? Faced with such timing decision what would you do?

Coding Horrors Community

Started by grumpyOldDude June 03, 2015 01:06 AM

11 comments, last by dubplatecode 8 years, 10 months ago

dtkaos

123

July 09, 2015 04:01 AM

I don't believe what you did was wrong, evil or egotistical. I believe you found a solution that needs to be refactored and while discovering complexity within your code.

I would refactor and keep it stupid simple (of course make sure you have comments for the code that doesn't speak for itself).

DekuTree64

1,170

July 09, 2015 03:58 PM

Did you learn anything useful in the process of writing the elegant version? If you always do the hacky/inflexible solution, that's all you'll be capable of. It's healthy to fight through a tough problem now and then. But you should choose which one by what will be the most educational to do right.

I assume you thoroughly understand your solution after that much time. Try looking at it from your former self's perspective... what could you have told him that would have helped him see what needed to be done sooner?

dubplatecode

215

July 10, 2015 02:57 PM

Keep it simple.

Figuring out a complex solution requires throwing code away, it takes longer to write elegant code, therefore its more time efficient to hack a prototype and "elegantify" after.

If the goal is a hobby project then consider what your acceptable level of elegance is vs completing the goal vs time. Sometime things take as long as they need theres no way around it but at least it gets done and move closer to the goal.

Sacrificing time: Elegance vs crude coding

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Sacrificing time: Elegance vs crude coding

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines