Decades ago British airways came up with an idea of a super-sonic commuter jet – The Concorde. At every stage of its development the project cost a lot more than was expected. It was clear at every point that a lot of resources were lost and a lot more will be lost before completion. So they could have pulled out at any point and saved a lot of money (in billions). But they didn’t. At any point the mentality was “If we quit now, all the money we have lost will be a waste”.
This is known as ‘The Concorde effect‘ (or ‘Sunk cost‘).
This is a very important effect to be aware of during the product design. Rationality dictates that we should not waste more money just because we have already sunk a lot of money in a particular project. However behavioural economics finds that humans have an opposite tendency when they face these situations. It is the result of their loss-averse nature.
Today is my turn to fight this irrational tendency in practice. In late October I introduced a tool called ‘Hammer’ in 3DTin. It lets you break a cube into tiny subcubes and delete some of them to create smoother gradients in your 3D pixel art. It sounds pretty cool. But it is not that essential as a feature. (All the sketches you see in the gallery on 3DTin front page… none of them required this tool). On the other hand, it has consumed overall 2 months of my development time. That is a lot, given that 3DTin has been in development for only 8 months so far.
The feature itself (subdividing cubes), didn’t take much long, but handling the resulting subcubes proved very painful – algorithmically. Every new feature I have added after that could have been implemented in less than half the time it actually took, only if I didn’t have to take care of the subcubes. If you follow this blog, you may have noticed that intermittently it goes quiet for a week or two. Guess what I am fixing in that time, instead of working on cool new features or usability bugs. Last week I came up with a much better algorithm to handle all the subcube mess, it was all working beautifully… except it made my CPU fan go crazy. At every step, I convinced myself that there should be a simpler approach to get this right and once I find it, it will all be a smooth sail after that. So I kept on. After all, I couldn’t abandon it, not after putting so much effort.
When you are involved in something very deep, you fail to see the big picture. That’s what leads to irrational behavior. It is clear now that the subcubes will continue to be a hurdle in all the features I plan in future. So I have decided to discontinue it, instead of falling into the irrational trap described above.
You will no longer see the hammer icon if you visit 3DTin now. I will keep the code for this feature in a branch and in future may be revisit it. Or may come up with a simpler approach to achive same effects that it provided.
You will see new features soon in near future, thanks to the newly simplified state of the code.
3DTin 