9

I looked at the tags, and noticed that there is no tag for data modeling.

Before I try to add one, (assuming that I can) I'd like to find out whether the community has already thought about such a tag, and decided against it. There is a tag, "dimensional-modeling", but that's more specific than data-modeling.

There are a lot of questions about data-modeling over in SO. And there are a lot of questions about database-design that might be better understood as questions about data-modeling. Some questions about data-modeling are being migrated over here. The criterion seems to be to leave the most amateurish questions over in SO, and to migrate the more sophisticated questions over here. Does that alter the advisability of having such a tag over here?

The most interesting case to me, over in SO, is a questioner who has little or no experience with data-modeling, but a fair amount of experience with object-modeling. Some database experts tell such a person to "forget everything you ever learned", but I think such advice is rarely helpful. OTOH, an object modeler is apt to think of data-modeling as a lame subset of object-modeling, but that's an oversimplification. But there is a surprising overlap between the two.

This is especially true when comparing OOA (Object Oriented Analysis) with conceptual data modeling (ER when used for requirements analysis). It becomes a lot less true when you comare OOD with physical data modeling.

So, should there be a data-modeling tag over here? Why or why not?

7

I think there is a risk of creating confusion due to tag overlap. Data modeling questions are usually tagged database-design. This is a tag that I spend a good deal of time on, so I see the modeling questions, the tuning questions and even questions that I would consider to be more architectural as opposed to design-oriented.

The question I would ask about a data-modeling tag is "How does it help to bring the right eye-balls to questions?" Do we really have that many questions of sufficient sophistication that we can expect people to choose data-modeling as a tag but not choose database-design? If every data-modeling question is also tagged database-design then does the tag help?

I don't object to a data-modeling tag - at least not at all strongly - but I think that tags need to be not only clearly defined, but that new tags especially need to be considered in the context of how they differentiate from what is already in place.

  • I think this answer is mostly correct, and I have flagged it as such. I have one quibble. A great many neophytes do not clearly distinguish between conceptual data modeling and database design. More specifically, they do not make any distinction between ER modeling as such and relational modeling as such. This is usually a symptom of a deeper problem, the failure to distinguish between analysis and design. Sad to say, many professional tutorials also fail to make this distinction clear. However, in terms of bringing the right eye-balls to the question, your point is on target. – Walter Mitty Mar 29 '13 at 9:58
  • 2
    If @WalterMitty is right in saying new users might not be aware of the dba convention of tagging such questions with database-design and @JoelBrown is right about the likely degree of overlap between the two, perhaps data-modeling should be created as a tag synonym for database-design? – chucksmash Apr 5 '13 at 15:08
6

Any time you think a tag can be well defined and helps clearly define the question scope, you're welcome to create it.

If a tag is widely generic then it probably won't do well as a tag because it'll apply to too many questions, and not help people understand the question scope easily enough.

5

Database design as a term is too broad and generic to be any actual good. It can mean all of conceptual design, logical design, physical design.

Many questions labeled "database design" are physical-level design questions, best performance questions, engine-specific questions, but the tag is really used for just any of them.

The data-modeling label seems to be consistently not used for that category of questions (rightly so), and that's a good indicator for filtering out the physical-level questions I have little or no interest in.

  • You make some good points. Nevertheless, I have come to the opinion that there is too much overlap between data-modeling and database-design to make the distinction between two tags useful. Some I'm not going to create a data-modeling tag. someone else might. – Walter Mitty Mar 30 '13 at 13:31
  • Just a small quibble in the wording above. I regard the conceptual data model as capturing what is discovered during the analysis phase, and not what is decided during the design phase. Thus, I would have said "conceptual analysis" or just "analysis" where you said "conceptual design". Many of the questions asked obscure the distinction between analysis and design as well as obscuring the distinction between conceptual, logical, and physical data modeling. – Walter Mitty Mar 30 '13 at 13:35
  • :-) with a bit of a stretch, I might say that conceptual design is designing the problem, and logical design is designing the solution :-) – Erwin Smout Mar 30 '13 at 16:14
  • I call it "analyzing the problem" instead of "designing the problem". But that's a terminology difference, nothing more. – Walter Mitty Mar 30 '13 at 21:04
2

I agree that there is a lot of overlap, but I also think there is clearly a use for this tag especially among beginners or people coming from the object oriented world like the OP mentioned.

As such I would suggest .

This clearly shows it is a sub-discipline and lets the questioners be more specific if they desire. Plus the modelers can follow the tag if they desire, and cut out some of the noise from the overly broad .

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .