What is it and how does an RTA work?

3

Recently I read some questions about RTA and noticed there are not too many explanations. Looking a little deeper into this subject I came across some conceptual issues, nothing about coding.

  

A real-time application (RTA) is an application program that works within a period of time that the user perceives as immediate or current. The latency must be less than a set value, usually measured in seconds.

  • What qualifies an application as being an RTA or not? Your time of execution?
  • An RTA requires a particular hardware platform to to work?
  • What would be the targeted development method (objects, models, tests, behavior) best suited to create an RTA?
  • For general platforms, there are more methods / languages (s) that optimize application performance?
  • What is the implication of the latency rate in an RTA?
  • asked by anonymous 07.12.2017 / 13:17

    1 answer

    2
      

    What qualifies an application as being an RTA or not? Your runtime?

    I do not like this idea of real-time application. An application can have parts in real time and parts that are not. If there is any part that is not the application it is not real time, but this does not mean anything. Specific routines need to be real-time.

    At specific points it is necessary that the response happens with a given maximum time. It is not a question of being fast or slow, it is being guaranteed that it does not take more than that. Of course, it is common for measurements to be low.

    Does an RTA need a particular hardware platform to work?

    Not necessarily, but there are platforms that can be difficult to maintain in real time. And there are some that make it easier.

      

    What would be the most appropriate method of development oriented (objects, models, tests, behavior) to create an RTA?

    This is all nonsense. You're looking at the wrong things. None of this has anything to do with real time . What happens is that very complex applications with a lot of abstraction becomes more complicated to analyze if the guarantees will be fulfilled.

      

    For general platforms, are there better methods / languages that optimize application performance?

    It has certain languages that require the use of a garbage collector or other mechanisms that can make execution times unpredictable, making it impossible to give the necessary guarantees. It may still be possible, but it will give so much work that it does not pay to use that language. So the guys usually use more C, C ++, Ada, Rust, and languages like that. Even these can not guarantee the real time, they only create less difficulties.

    Apart from this, there is nothing specific about mainstream languages.

    It has operating systems that help RT, others hamper.

      

    What is the implication of the latency rate in an RTA?

    When latency exceeds established linite, you lose an opportunity you could not. It can be a frame of a game that delays rendering and impairs gameplay, may be a stock-and-stock-market control that accidentally ends up privileging one operation over another that lags behind the application, it may be a simulation that produces distorted results, or it may be an airplane control that delays and causes an accident to run out of the moment it would need, or still not pick up an error that can only be identified at a time or even create a race condition that would not happen if time were respected, and to close one of the countless examples can create a disruption in something that you need to maintain a linear sequence.

        
    07.12.2017 / 15:01