TIBCO EBX Room for Improvement
Lead Data Modeler - Enterprise Data Strategies and Services at a financial services firm with 10,001+ employees
It's a pretty steep learning curve, I find. And not really fluid and flexible. There is some graphical rendering of the workflows, however, you can't really develop them in terms of the graphical picture. Whereas a lot of BPM-type tools will give you that kind of capability.
The workflows need improvement. You need to develop, kind of conceptually, what you want. It's basically a web app generator, so there's a lot of magic under the covers. When you're trying to promote the changes through a version control system, it's hard to know what to expect as far as all the content. For example, if we were building and writing Java code, we would know what's changing. However, due to the fact that we're just putting in models and embedding some business logic in the models and such, it generates a web app, a job of whatever. It generates XML and some other stuff. And that's XSD. Then when you go and say, "Okay, let's push these changes," in Git or in Eclipse, et cetera, it's tricky to have a multi-developer environment where you're not stepping on each other a little bit. You're not as aware of the repercussions of your design changes.
The documentation needs improvement. It would be helpful to have more during implementation, for example. It would help make the initial setup more straightforward.
In a workflow, you can't set default values for certain columns, which would be nice.
If you're handy with Java, you can create your own services and such and do something there, however, it should be out-of-the-box functionality. If you have a generic system, you should be able to say, "Hey, this structure supports A, B or C." Yet, if you launch it, if department A launches it, assume that certain values are set to their preferences. As department D launches assume that the department is B and you know, certain values are set, et cetera. Otherwise, everybody comes in generically, and then they have to know more than they want to know.
There's this thing called replication. You could replicate the XML database to SQL Server on Oracle. That replication doesn't happen if you use certain features of the product. For example, with one of these features, you can do calculations or calculated fields. You could say it's X, then do a sum, et cetera. If you have a calculated field, you're not allowed to replicate it. It would be better if they could allow the replication, even if maybe they have to limit the functionality to those columns.
There's an item called inheritance. You could say, I don't want to, if I asked for your name, I don't want to ask for your name three more times. So when I get to another file on the table, it's already there. It will carry information by inheritance, so you're not going to enter it wrong three different ways, however, once you have inheritance, you can't replicate. It would be better to ignore the inherited fields, just nix those columns, yet allow the table to replicate. Then you can have an SQL server go in and read the data in a relational way, which is very helpful to make it acceptable to developers and business analysts, et cetera.
Sometimes it is a little difficult to find an architect for certain developments. You kind of have to be like technicians to work with this product if you want to make some specific development. So perhaps it could be simplified. The design should also be easier and more visual.View full review »
In terms of improvements that this solution needs, it should support big data.View full review »
Head of Architecture at a insurance company with 5,001-10,000 employees
The support isn't great and takes too long. And the pricing can be improved.View full review »