Although the domain layer meta model is very simple and at first glance creating a customized model seems easy, identifying appropriate entity types and functions is not. Difficulties will (most likely) arise in understanding what entity types exactly are. To avoid these difficulties, it is recommended to consult again the definition of the term entity type: “...Entity types describe the character of the information that is needed to perform a function or that can be made available after the function was performed.” It follows from that that if, for example, a medical report is to be modeled as a physical real life object, it should be taken into consideration that such reports also contain information about patients, cases, diagnoses, etc., which are usually entity types of their own. Certain problems that are to be modeled might call for visualizing this kind of relation (as ‘part_of’ relation). Only these relations represent successfully the complex character of reality. For reasons mentioned above, we recommend using reference models on the domain layer (if available), the archive reference model, for example.

2. Is it possible to graphically visualize organizational units (e.g. modeling organizational units in combination with the associated functions)?

In the interest of keeping the graphic model clear, it is presently not possible to graphically visualize organizational units.

3. Why can message types, document types and dataset types, i.e. all representation forms of entity types, be assigned in the dialog window Entity type/Properties/Representation forms on the domain layer although they are actually part of the logical tool layer?

Message types, document types and dataset types are part of the logical tool layer. Only representation on the logical tool layer is controled using the property dialog of an entity type. Also related to this pocedure is the possibility to create new message types, document types and dataset types to facilitate the modeling if the entities needed have not been modeled so far.

4. What is the drop-down menu ‘Master-DBS’ in the dialog window Entity type/Properties/General on the domain layer for?

By using the drop-down menu ‘Master-DBS’ the user can select which database system/document collection is to serve as master for an entity type (‘has_as_master’ relation in the meta model). In the dialog window, the term ‘master’ means that the assigned database system/document collection contains all current data at any moment. Other database systems/document collections that store the same entity type have to synchronize their data with the Master-DBS via communication links between the associated application components in case any data is changed, added or deleted in order to ensure data integrity.
(This is especially a problem with data collections!)

2. Logical Tool Layer

5. How can I model a database that is not associated with any of the application components?

What is probably meant here is a database system. Potentially, a database system itself can be an application component. In this case, the first step is to define an application component that acts as a metaphorical cover. The application component is then assigned a database system as well as component interfaces through which other application components can access the database.

6. Is it possible to create several database systems for one application component?

According to our meta model, an application component cannot be assigned several databases. Since all application components of the 3LGM² Tool have only one database system each, our experience shows that such an option is not needed anyway. However, it is certainly possible for an application component to access several database systems. In this case, these database systems and the application components they are assigned to must be modeled first (see also FAQ 5.). Component interfaces and communication links have to be created next to ensure access.

7. Are there any modeling standards for paper-based application components on the logical tool layer?

We are currently developing modeling standards. In the meantime, further information can be found in the following article: Modeling Hospital Information Systems (Part 2)

8. When is an interface a user interface and when a component interface?

Whenever an interface is used to model a communication link to another application component, it is a component interface. A user interface supports communication between user and tool and is more about software ergonomics.

9. Why is the message type linked to an entity type in the dialog window Entity type/Properties? Should not the message type rather be linked to the dataset type/the document type (e.g. message type ‘ADT message’ – dataset type ‘patient details’ instead of the entity ‘patient’)?

Our meta model offers three forms of representing an entity type on the logical tool layer: as a dataset type, as a message type or as a document type. Dataset types describe how the information is stored in a database system. Message types describe the communication between two computer-based application components. Document types describe how the information is stored in a document collection as well as the communication between two application components, of which at least one is paper-based.

10. Is the Master-DBS the database management system displayed in the model browser?

No. See also FAQ 1.4.

11. When can it be necessary to model paper-based component interfaces (e.g. when paper-based documentation enters a computer-based application component – physician order form)?

Users often want to model media cracks of data transmission between computer-based and paper-based application components. In this case, the following questions have to be answered:

Is data captured on paper forms and then integrated manually into a computer-based application system?
Is data printed from a computer-based application system and put to further use on paper?

Please note: As users are usually not modeled in 3LGM² models, the use of a computer-based application component is also implicit. Therefore, printing, reading and destroying a document is usually not modeled, because the user could just as well have read the document on screen. Yet if the scenario is to be modeled (if unnecessarily high paper consumption is suspected, for example), it can be done by modeling a paper-based application component called ‘Operation’ with a document collection called ‘Wastebasket’.

12. As I see it, interfaces cannot do both transmit and receive, yet interfaces in the meta model and the 3LGM² Tool do exactly that. Why?

Our meta model allows interfaces to do both transmit and receive. This is hard to understand at first, since real life interfaces are seldom used for both purposes. However, the likely need for more abstract modeling should also be met. If, for example, the task is merely to model two application components that communicate with each other bidirectionally, each application component is assigned an interface and these interfaces are then linked in both directions. This is an option for modelers pursuing different aims and creating different levels of abstraction in their IS.

13. How can my database system be assigned its Master-DBS on the logical tool layer?

This is not possible. See also FAQ ‎1.4.

14. Only document types offer ETNT combinations, dataset types do not. Yet how can I still model these in the case of two computer-based application components?

It was our intention to exclude an option for ETNT combinations in dataset types because computer-based application components exclusively communicate via message types. Dataset types only represent how entity types are stored. By contrast, in paper-based communication situations document types are used for both communication and storage of entity types. Therefore, ETNT combinations (which were actually to be called ETDT combinations) are required.

15. Why are software products assigned to functions and not to application components in the dialog window Software product/Properties?

The equivalent relation in the meta model is ‘can_support’. It enables the user to parameterize which of the functions a software product can support are actually supported after the installation. The ‘can_support’ relation theoretically helps to determine whether an application component based on a certain software product should support more functions – to reduce the level of heterogeneity, for example.

16. Why are application programs assigned to software products and application components? I thought that application programs were adapted software products (specific parameterization, etc.).

Application programs control application components (‘is_controlled_by’ relation in the meta model). An application program is an adapted software product (‘bases_on’ relation in the meta model). To simplify things, a software product can be assigned directly in the properties dialog of an application component, but is actually assigned to the application program controlled by this application component.

3. Physical Tool Layer

17. Why is a subnet not simply called a net? Why are there subnets if there is no net?

There is no detailed explanation as to why. In most cases, nets/subnets only make sense in connection with subnets. The information that all physical data processing components of a model belong to a single net is almost always useless.

4. Inter-layer Relationships

18. Why are application component configurations linked to organizational units??

Very often it depends on the organizational unit which application component configuration supports which function. This relation between the three classes function, organizational unit and application component configuration was introduced in order to be able to model the question which application component configuration is used in which organizational unit.

19. Is it possible to assign database management systems to data-processing component configurations?

Since the meta model does not provide for that, neither does the 3LGM ² Tool. When a database system is assigned to an application component, assignment to a component configuration is implicit. Maybe you could give us a short example to motivate us to adjust the model and the tool. We also might find another way of modeling this scenario with presently available means.

5. Miscellaneous

20. What do these terms mean: ‘is_part_of’ relations of functions, entity types, organizational units, application components and physical data-processing components?

Using the 3LGM² Tool

1. Domain Layer

2. Logical Tool Layer

21. Does DBS in the properties dialog of the database management system refer to the database system?

Yes. (For a better understanding the tool should be adapted in that point!)

22. Are application programs named automatically and, if this is the case, according to which rules?

Application programs are not named automatically and they are not created automatically either. An application programm can only be created by entering a name for the application programm in the properties dialog of the application component.

23. On the logical tool layer, why is it not possible to create new entity instances in the model browser for every single element?

There are model elements in the 3LGM² that are connected to another model element (on the logical tool layer this is mostly the application component) by a dependency, which means that those model elements may only exist if the corresponding application component exists. For that very reason it makes little sense to to have the possiblity to create entity instances in the model browser for those model elements.

3. Physical Tool Layer

24. Why is there no graphic representation for network types ?

An appropriate visualisation concept for network types has yet to be found.

4. Inter-Layer-Relationships

25. How can I create the application component configuration to a function?

By using the tab ‘Application Component Configuration’ in the function’s properties dialog. Application component configurations always refer to an organizational unit. Therefore the first step would be to choose an organizational unit in the window’s left part. Then a set of application components (=application component configuration) can be selected from the window’s right part and be assigned to the chosen organizational unit using the arrow keys. If there has not yet been any organizational unit assigned to the function, this is still to be done on the tab before proceeding to anything else.

(Tip: If the intention is to neglect the organizational units but still create application component configurations, an organizational unit ‘hospital’ could be created (as a kind of dummy) and all these functions can be assigned to it.)

5. Miscellaneous

26. Is there a possibility to enlarge the modeling canvas? The drawing of bigger models could require exactly that.

The size of the modeling canvas can be increased or reduced via the view settings by shifting the fader to the right or to the left.

27. What sense does it make to have the possibility to create elements in the context menu that specify other elements (e.g. software product…) and that can’t be represented graphically, but are only listed in the model browser which makes it hard for a beginner to assign them later on?

The instances of a model element created in the model browser are not only shown in that very browser but they are also to be found in the properties dialog for further modeling. The possibility to create instances via the model browser’s context menu expands the modeling strategies. A modeler for instance would have no problems with entering all software products used in his organization that he knows of at the beginning so that they are at his disposal during the subsequent modeling of the application components.

28. Which contents are hidden in the context menu “internal (verification, interactive…)”?

The significance of the instructions to be found under “Internals” is restrained to the development of the 3LGM² Tool. They are of no importance to the tool’s use.

Last updated: 24-Apr-2019 11:44:04

	Deutsche Website ImpressumDatenschutz
		Institut für Medizinische Informatik, Statistik und Epidemiologie