Kalido Business Information Modeler - everyone should have one

I posted about Kalido Business Information Modeler back in February, but now it's gone GA and Philip Howard reviews it at IT-Director. He is pretty enthusiastic and says "everyone should have Kalido Dynamic Information Warehouse and they should certainly have Universal Information Director and Business Information Modeler too."

The only negative he comes up with is "the relatively limited number of platforms that Kalido's software runs on: one would like to see it on Netezza for example, or Teradata".

Shared Business Vocabulary

Following on from my active metadata post yesterday, Mike Ferguson at Dataflux says Shared Business Vocabulary (is) needed everywhere. By "shared business vocabulary" he means metadata, as this earlier post explains. I don't particularly like his phrase - metadata is fine by me - but maybe he's right and we need to use a less loaded (and misunderstood) term for it. And particularly in the warehousey world "metadata" also refers to the dynamics of data (when was it loaded, how long did it take) so there's scope for a better word or phrase.

Whatever that word or phrase is, please please can it not be ontology ....

Active Metadata

Yesterday I moaned about poor use of metadata. Today I'd like to point to an example of the way metadata can be used actively. Andy Hayler describes Kalido's Business Information Modeller, which allows you actively to control and when necessary reshape the data warehouse by changing metadata; you can generate a BO universe, or deploy metadata to Cognos BI tools (more to come, apparently). You can (apparently) even undo the metadata changes, and see your warehouse as it would have been; kind of like Oracle's flashback recovery, but for data warehouse structures.

See Kalido's podcast, screenshots and other resources.

2008 Data Management Predictions from Dataflux

Mike Ferguson posted these over a week ago. Key thoughts to take away from my point of view:

• Information and data architects will continue to be in demand
• Companies will need to invest again in data modeling tools and in data modeling skills
• Holding this metadata in spreadsheets is no longer acceptable.

It is depressing to me that so many projects keep their business critical metadata in Word documents, Excel spreadsheets and Visio drawings. Some developers prefer to rely on their IDE - keeping DDL definitions as SQL in text files. That protects the definitions (at least you can find them quickly), but it doesn't get the maximum value out of it.

CASE tools have been around since the mid 80s or before - Oracle's SQL*CASE, now better known as Designer, was under development when I joined in '86 - so how come they are used less and less? They did get a bad name for encouraging complex, expensive and ultimately useless corporate data models - and we're glad to see the back of those, and the ivory towers they came from - but they can still be very helpful in defining and developing the metadata we need as a basis for system development.

I wonder whether the main problem is that many CASE tools are simply too expensive and/or too closed; they just can't cope with all the different kinds and layers of metadata we would like to throw at them, and they don't integrate well with all the other development tools around. Look at Designer - it's been more or less static for the last 10 years, and other Oracle products barely take any notice of it. No wonder it's slowly fading away.

Perhaps it's time for Oracle to get a grip and provide some common repository / metadata management for use across all its myriad of tools? Or for a small vendor or OSS project to take up the challenge? Let me know if you've already found the tool that can pull together ERDs, schema models, UML process diagrams, Discoverer EULs, a BO universe, Warehouse Builder or ODI transformation definitions and all the other kinds of development metadata that projects deal with every day.