I agree there is no such a standard for ER modeling. In fact, some organizations have their own standard, depending on their needs and knowledge of their teams. On the other hand, we have a myriad of coding standards, like CamelCase and other good practices. I believe it's a matter of how important is to have a modeling standard in order to enhance value to software products deliveries, even more because of the agile ways of developing.

Ok, Fernando Aparecido Nogueira , thank you for answering. That's basically the state of affairs. But what about the motivations? Why there is no such standard yet?

The OMG, for example, have a lot of standards for almost every level of abstraction. However, the Information Management Metamodel Vol. II has an approved proposal for an ER metamodel since 2009. Business analysts use the ER diagram as a tool for requirements engineering. Almost every databases book known, has a chapter for the ER model. Every introductory database course has an obligatory topic for ER modeling.

But this is a fact: several authors claim that ER model is very popular, and the software industry use it.

So, why in Peter Chen's name there is no such standard yet?

That's a good question. OMG standarizes UML which can be used for describe databases like ER. On the one hand, two standards for the same requirements is too much. On the other hand, hitoricaly ER was first so in my opinion should be standarized a long time ago.

Thanks for answering, Dominik Tomaszuk. Database designers are using UML class diagrams as a mean of expressing an ER scheme. Even more, many database book classics have introduced UML class diagrams for this purpose.

However, UML is a language created for the Object Oriented paradigm. And because a language is syntax and semantics, the OMG has standardized both. So, what is happening is that we are forced to adapt UML class diagrams to ER modeling, resulting in a variety of "adaptations" as every author recommend different strategies. Therefore, we can conclude that OMG has standardized UML as an Object-oriented system analysis and design language, but we are still lacking of an ER approach standard for database design.

The UML Class Diagrams are used for DBMS as well. Moreover UML is used for business modelling now as well, not just for Object Oriented design. Indeed, UML originally was developed for the Object Oriented paradigm, but now used in many other areas, including for ER modelling.

That's true, Youry Khmelevsky, UML Class Diagrams are used for database design as well. And as I said before, "many database book classics have introduced UML class diagrams for this purpose". And, as I also said before, "what is happening is that we are forced to adapt UML class diagrams to ER modeling, resulting in a variety of "adaptations" as every author recommend different strategies". So, we've ended up in the same situation.The reason is simple: UML Class Diagrams, although inspired in former ER diagrams, were not meant for ER modeling. So we have to use stereotypes, profiles, etc. Or worst, we have to discard some OO notations like composition or aggregation. Also, we have to omit the methods section of a class box. That is like speaking Japanese with Chinese characters (I don't know Japanese nor Chinese, so I don't know if that's possible, but I imagine that is not straightforward nor a recommended practice). Of course, you may model your business from an OO perspective, and that is OK. BUT, ER modeling approach is not standardized yet.

Remember, a language is syntax, semantics and pragmatics. And in that sense, we don't have a consensus nor a standard about ER modeling.

Why? Is it possible? Is it viable to have one?

UML 2.0 allows own/new notations and stereotypes. Many companies use it for own purposes. I personally don’t see issues with Class Diagrams for ERD, but it can be extended to have “classical” ERD.

Thanks for answering, Youry Khmelevsky . This question is not about why-not-UML, is about the motivations for not having an ER model standard yet.

UML Class Diagrams for ER modeling is widely used in academic teaching and industry. That is not in debate.

However, UML class diagrams were not created for ER modeling, although we can extend it or adopt a discipline in using those diagrams while we are thinking in an ER model. But the different origins have consequences. We have to add stereotypes (which ones?) or properties to the same thing (identifiers, for example), modify what a role name means, be sure we are using domain-specific entity classes (because an object could be everything, but an entity class is a collection of things of significance in the domain being addressed), use symbols selectively (using some UML symbols while not using others). As we can see, there are decisions about the modeling language that are not standardized. Also, from a teaching perspective, we have to explain to students HOW to use and think about UML class diagrams for ER modeling. This is not necessary if, for example, we use a notation like IDEF1x or Barker-Ellis. But there are a lot of ER notations, and that's one of the reasons why we need, in my opinion, a standard ER model.

I recommend a book from David C. Hay, UML & Data Modeling: A Reconciliation. He recognized that there is a difference between UML community and ER community, and proposed a practice for using UML class diagrams for ER modeling. But a brief reading of Connolly, T. M., and C. E. Begg. Database Systems: A Practical Approach to Design, Implementation, and Management. 6th ed., Pearson Education, 2015, shows different recommendations for reconciling UML class diagrams with ER modeling. Also, CASE tools have different profiles for ER modeling with UML class diagrams.

In conclusion, UML is standardized, but ER model is not.

I agree with you. Probably we need an official/standardized UML Class Diagram notation for ER modelling, but it will take time.