LabVIEW AF is a fantastic tool to create multi-threaded messaging applications the right way. It promotes best practices and help write big yet scalable applications.
I feel however that there is not enough included in the box 🙂 When looking at many other industry implementation of the actor model e.g. Scala Akka, Erlang, Elixir, microservices in many languages etc. you find that the frameworks usually include many more features. https://getakka.net/
I understand the decisions behind the design for AF yet would like to start a discussion about some of these.
I would like to see the following features being included into AF:
1. Ability to Get Actor Enqueuer by Path through hierarchy of actors e.g. /root/pactor/chactor could be used to get the enqueuer to chactor. This would probably require actors to have unique paths pointing to them. Having a system manager keeping all enqueuers on local system is not a bad thing. It is a good thing.
2. Ability to seamlessly communicate over the network and create your own queue specifications. For example actors should be able to talk between physical systems. Using the Network Actor is not a good solution, because it is simply to verbose and difficult to understand. The philosophy of AKKA should be embraced here "This effort has been undertaken to ensure that all functions are available equally when running within a single process or on a cluster of hundreds of machines. The key for enabling this is to go from remote to local by way of optimization instead of trying to go from local to remote by way of generalization. See this classic paper for a detailed discussion on why the second approach is bound to fail."
3. Improved debugging features like MGI Monitored Actor being inbuilt. https://www.mooregoodideas.com/actor-framework/monitored-actor/monitored-actor-2-0/
4. Included Subscriber Publisher communication scheme as a standard way for communicating outside the tree hierarchy. https://forums.ni.com/t5/Actor-Framework-Documents/Event-Source-Actor-Package/ta-p/3538542?profile.l...
5. Certain templates, standard actors, HALs, MALs should be part of the framework e.g. TDMS Logger Actor, DAQmx Actor. Right now the framework feels naked when you start with it, and substantial effort is required to prepare it for real application development. The fact that standard solutions would not be perfect for everyone should not prevent us from providing them, because still 80% of programmers will benefit.
6. Interface based messaging. The need to create messages for all communication is a major decrease in productivity and speed of actor programming. It also decreases readability. It is a better with the BD Preview in Choose Implementation Dialog in LV19, but still 🙂
7. This is more of an object orientation thing rather than actor thing but would have huge implications for actor core VI, or Receive Message VI. Please add pattern matching into OO LV. It could look like a case structure adapting to a class hierarchy on case selector and doing the type casting to the specific class inside the case structure. You could have dynamic, class based behavior without creating dynamic dispatch VIs, and you would still keep everything type safe. https://docs.scala-lang.org/tour/pattern-matching.html
The natural way for programming languages to evolve is to take ideas from the community and build them into the language. This needs to also become the LabVIEW way. I saw a demo of features of LV20 and I am really happy to see that many new features were inspired by the community. Lets take some ideas about actor and add them in.
I wanted to share my view on the direction I think AF should go. I see in it the potential to become the standard way to do big applications, but for that we need to minimize the barrier to entry.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Please post each idea separately so they can be evaluated individually instead of as a group.