Why do I mention all of this here?  Simply to give you an idea that data modelers do not easily switch tools.  A vendor has to offer most or all of the features already in their existing tool of choice as well as significantly more functionality and ease of use in order for someone to switch. 

ER/Studio 3.0 is the most recent incarnation of Embarcadero's data modeling tool.   I first looked at this tool a few years ago when it was in its 1.0 release.   At that point it was an entry into the market which wasn't worth noting when compared to the likes of ERWin and S-Designor.  They had asked me to look at the 2.5 release of ER/Studio.  At that point, it was apparent ER/Studio had made some great strides.  But, it still wasn't up to much of the functionality that I relied on with S-Designor and I really didn't have much time to give it a worthwhile run through.  Version 3.0 is now out and as usual, Embarcadero is after me for feedback and what I think about it.

As a side note, Embarcadero has been an extremely pleasant company to work with.   They will listen to everything you have to say whether good or bad.  They are also very responsive.  I generally get responses to every e-mail I send within an hour or two.  Tech support usually gets back to me within an hour or two.  From working with Embarcadero, I know this isn't preferential treament.  All of their customers and potential customers get treated that way.  If you have an idea for improving a product they are all ears and you will very likely see that feature implemented within a version or two.  This is definitely a company with its act together across the board and customer service is top notch.

I have been very hard on ER/Studio in terms of functionality that I want to see.   To the credit of Embarcadero, they have listened to each suggestion and evaluated it.  Many of those are in version 3.0 or slated for future versions.  I can say they still don't have me completely convinced at this point, because I am still pretty tied to S-Designor mainly because of the graphical capabilities.  But, it is getting very close.  I'm currently switching back and forth between the two modeling tools to take advantage of features of each.   (We'll have to see if they get me to completely convert with the next release.)

If you are evaluating data modeling tools or not currently using one, ER/Studio rates a very strong consideration and I would recommend adding it to your toolset.  Version 3.0 has moved it up into the arena where it can compete favorably with its chief rivals of ERWin and PowerDesigner.  It finally offers significant and compelling functionality to help in your modeling efforts.  I won't tell you it is all roses.  It is still a little rough around the edges, but it does a very solid job of data modeling and those rough edges are being worked on as you read this.

Any data modeling tool you pick up will have a zillion features.  I am not going to even attempt to cover some of those features, simply because using some of them gets into data modeling techniques which are better handled in another article.  Some of the features are available to accomplish specific functions within relational design theory.  That is another topic I won't get into.  Even taking some of that stuff out, ER/Studio offers an overwhelming array of features.  I will attempt to cover as many features as possible.  I plan on updating this review as I get more familiar with the capabilities and hopefully, I will be able to eventually cover all of the features available within ER/Studio.

One other thing that needs to be addressed before showing off this product is the question that comes up very often: "Why do I need a data modeling tool?"   The answer to that question is actually pretty ambiguous.  It is along the lines of: "You can tell me once you have actually picked up a modeling tool and built a database with it."  That doesn't really help you out though.  If you are designing small databases with a handful of tables in them, a data modeling tool is probably overkill and would be pretty hard to justify the cost.  But if you are putting together complex or larger databases, a data modeling tool will pay for itself so many times over it is a very simple choice.

What a modeling tool gives you is a very powerful, graphical, and easy to use interface to designing a database.  You literally create your database via point and click.   It also gives you a very good validation method and allows you to automatically enforce certain relational constructs.  The database designer gets a nice view of the database being designed and a very simple way to make changes.  Once a table or relationship is implemented in a database, it is pretty difficult to modify.  The data modeling tool alleviates much of this rework by allowing you to modify your schema simply and painlessly.  Because of the ease of use in the interface, creating databases can be reduced from months to weeks, weeks to days, and days to hours.  If you put it in terms of ROI, it is very simple.  I usually bill out at a client at around $100/hr.  Each of the top three data modeling tools (ER/Studio, PowerDesigner, and ERWin) run around $2000.  If a modeling tool saves me 20 hours on a project, it has paid for itself.  I can pretty much guarantee you will save that much time within the first week or two. Data modeling tools have saved me so much time over the years, that I will not even consider designing a database without one.  In addition to the graphical design saving time, every modeling tool also generates code for you.  This means that you can graphically design your database and then have the tool generate all of the DDL required to implement it.  Everybody always complained about using triggers to implement RI.  They said the biggest drwaback was writing all of that code and the possibility of mistakes.  With the code generation features of a modeling tool, you write absolutely no code.  You simply select the type of relationship it is via radio buttons.  The tool will the generate the proverbial 10 miles of trigger code for you.   Very simple and extremely painless.  If you are working in heterogenous environments, it gets even better.  You can design a database with one target DBMS and simply by changing a drop down selection, you now have a database that conforms to a completely different DBMS.  The modeling tool does all of the changes internally.

In addition to being able to create new data models, ER/Studio is a round trip data modeling tool.  This means it has the capability to reverse engineer an existing database, make changes, and then synchronize those changes back to your database.   This provides a means for modifying existing databases and also migrating them to another DBMS.  Support is provided for 27 DBMSes which should satisfy just about all of your modeling needs.

For the purpose of this review, I utilized the Northwind database that comes with MS SQL Server 7.0.  This was not only for convenience, but also to demonstrate that it is compatible with 7.0 and you will not be forced to upgrade immediately.  The screen shots you will find are done at an 800x600 resolution for readability purposes.  In real life, I model on the largest monitor available.  My preferred system right now is dual 21" monitors linked through Windows 98 which provides enough screen real estate for even the largest data models.

Within this review, there is certain terminology I will use.  ER/Studio makes logical and physical data modeling virtually transparent from an appearance perspective on screen, so I will use entity and table interchangeably.  I'll also use attribute and column interchangeably.  The screens and editors look almost identical for each.  For those purists, you use entities and attributes in logical modeling and tables and columns in physical modeling.   I do have to confess that I have never put together a logical model in my life.   Every time I have tried, it winds up looking like a physical model.  Because of this, I always go directly to a physical model and start from there.  People will argue both ways and for those who have participated on the MS SQL Server mailing list know this can turn into a huge debate as to what is the "correct" way.  My stance is that the "correct" way is what works best for you.  The end goal is the same and that is to quickly and efficiently produce a database that meets the business requirements for your application.  Please no message trying to convince me othewise.  After putting together a few hundred databases, including one that was over 3500 tables, I am not changing simply because I don't think that way.  This is one of the biggest things I like about ER/Studio.  It does not impose one or the other on you.  Nor does it impose a specific order.  I can start inside of my physical model or inside of my logical model and just start designing.  The tools are in place to keep the two synchronized.  When you create a new model, you only have a logical model.  But all you have to do is drop a dummy entity and then generate a physical model.  From there, you can work in either one.

ER/Studio 3.0 1 2 3 4 5 6 7 8