Today, one of the people in our team, after reading an email I sent about the different types of robots we have in so many other production environments, asked me: “Where do robots go when they die?
Indeed, he asked with a half-smile in a clear humorous tone. However, it happened to me as with almost all the different questions I am asked: it made me think. And on this occasion, the thread of thought led me to another question: – “What is the life cycle of a robot?
“What’s the life cycle of a robot?”
Everything that has a beginning has an end
A robot is a piece of software, more or less complex, more or less extensive, more or less modified since its construction, but like any piece of software, it will end up “dying”. And like any biological, mechanical or computational element, death will be more or less honorable and happy depending on how it has taken advantage of the life it has enjoyed.
In terms of software robots, this use of the time granted can be measured in some very simple parameters. We can compare the life of a robot with the life of a person.
From the conception and construction of the robot, we can know if a robot has been “born well” if the number of hours of real development is lower than the estimated development time.
Once passed this first phase, the birth, we arrive at the training period, which is the time invested in properly adjusting a robot already built to pass from a development environment to the real environment for final execution. This time is also a measurable element.
In this phase, we may find that the production environment is significantly different from the development environment, both in terms of performance and volumes of information or even versions of applications! If an extensive adjustment is not necessary, we can say that the robot “is growing well”.
Already when the robot is an adult, and must “earn its bread”, we can measure more and effectively, as for example:
- Number of executions carried out in production
- Number of items processed correctly
- Total execution time
- Temporary and economic savings obtained
- Number of incidents
- Number of versions deployed
- Hours of production support required
This is the moment when more metrics can be collected, and not only from the point of view of maintenance but also from the point of view of a business. In the Jidoka console, you can find metrics about the execution of the robot and create follow-up reports.
When it’s about time to deploy a robot, it’s also possible to take stock of whether the development of the robot has resulted in some kind of improvement or new functionality in the platform. In that case, the robot will be worth it, at least for me.
It is also necessary to consider the improvement of the processes assigned to it. A good robot, one that has fulfilled its duty, would not surprise me if it had managed to change the processes it has performed for the better.
We can also evaluate the pleasure or satisfaction (perhaps this word is more appropriate) that the construction and maintenance of the robot caused in the developer over time. If the developer is happy, the robot may have earned a presence in the “hall of fame” or “robot sky”.
Note: “The sky of the robots” is a piece of cork that we have in our office where we click on what we have built, and which has been most gratifying to us.
Undoubtedly, this way of seeing the life cycle of a program may seem at least something “different” (for being kind), but we should not always see things just as they are. There are times when we can afford a small luxury: using a filter to laugh a little.
CTO at Jidoka.