Vorto Meta Model Discussion [message #1725453] |
Thu, 03 March 2016 15:19 |
Henning Groenda Messages: 11 Registered: July 2009 |
Junior Member |
|
|
Hello,
when comparing Vorto and SensIDL a few questions came to my mind:
* Is there a difference between displayname and name? A single general type named element would ease consistent access. Its not class-specifc anymore and i could see no class that has both, a displayname and name.
* Property: multiplicity of type EBoolean? I would have expected bounds.
* Why Reference from FunctionBlock to Enum and Entity instead of to the more general Type element?
* Why is there no separation between Type and a possible TypeModel, which would be the consistent modelling inferred from the same concept at FunctionblockModel and InformationModel?
* ConstraintIntervalType - Why is Interval in the name? Why are strlen, regex and mimetype in there? Those are not constraints. Its unclear whar regex could mean.
* The semantic difference and therefore use of the alternatives Event, Fault, Status and Configuration remained unclear. Can you point me to a explanation?
* (One question from my colleague Chris:) Is there a reason for not reusing Property and related element definitions for Operation (+ Param, RefParam, PrimitiveParam, ReturnType, ReturnObjectType, ReturnPrimitiveType)? Indeed, they seem very similar when looking at the concepts and structure.
I am curious to learn about the meta model, your solution and decisions and looking forward to your reply.
Best regards,
Henning
|
|
|
Re: Vorto Meta Model Discussion [message #1725475 is a reply to message #1725453] |
Thu, 03 March 2016 17:32 |
Olaf Weinmann Messages: 29 Registered: May 2015 |
Junior Member |
|
|
Hi Henning,
these are very good questions!
* The name of a model is part of the unique identifier within the Vorto Repository (namespace, name, version). The displayname defines the appearance of the model in the IoT Toolset. You're absolutely right that this may lead to an inconsistence. We should think about harmonizing that.
* I would say that bounds can be considered as constraints. But the multiplicity in general should be discussed since we currently do not allow multi-dimensional arrays out of the box. Suggestions and ideas are very welcome!
* Your question related to our Functionblock references: This mechanism could be improved, I agree. In addition to your suggestion Functionblocks should be included into the reference mechanism. I'm pretty sure that we'll need a way to support inheritance in the near future.
* You're absolutely right we should remove the word "interval" within the constraints enum.
* The existence of the groups "fault", "status" and "configuration" has historical reasons. We'll remove these groups but add a more flexible solution for specific groupings.
* There is no reason for not reusing Property for Param, RefParam,.... etc. The reason for the difference is, that in the very beginning of Vorto we generated the meta model out of the grammar of our DSL. Of course we need to harmonize these parts of the meta model.
Thank you very much for asking these questions and highlighting these points. Please let me know if you have further questions!
Cheers,
Olaf
|
|
|
Powered by
FUDForum. Page generated in 0.21659 seconds