Recently in BI Technology Category

Maya from Autodesk is a high-end 3D modeling and rendering tool. Many commercial productions of computer animation, such as Ice Age and Monsters/Aliens, use this tool. And, many professional artists spend thousands of hours immersed in this tool.

I heard that Buzz King, a colleague at CU Boulder, was teaching a class on Computer Animation using Maya as the primary learning vehicle. I thought that this is a class would be good for me to understand Maya and the latest techniques in 3D modeling.

It has been a maturing experience. Let me share with you the maturing stages experienced and my reflections about Business Intelligence tools.

1) Complexity Kills Usability

My first impression is "you got to be kidding me!" No product could be this complex and still be commercially viable. The user interface has so many icons to click and so many levels to those icons. Most normal humans would take a look at this tool and immediately throw up. However, Buzz is a friend so I did attend the second class.

2) Rich UI Requires Patience

I eventually came to terms with the UI complexity, but only when I could achieve amazing results in 3 or so clicks. So, I thought... Maybe this complexity does have value and just requires a bit of patience. I quickly realized that learning Maya requires huge amounts of patience...with the tool, with yourself, and with your sore butt.

3) Clicking Hinders Learning

I learned next that your patience must be mixed with curiosity, or Maya degenerates into boredom and frustration. Behind all this complexity is an amazing tool! My instinct was then to click on all the icons. I thought, "There must be some interesting reason for this oddly shaped icon." I soon realized that undo does not undo everything and that actions are constrained by mysterious pre-conditions. Either way, the results were weird, or nothing happens, which is even more frustrating. I concluded that you really need to know what the icons mean. Think first; then click. The ready-shoot-aim approach does not work!

4) Concepts are Mind Bending

I next dug into the basic concepts embedded in Maya. Drawing pretty images is a small part. Understanding the rendering process and extending that to animation as a creative production is mind bending. My advice is... "Free mind, twist 67.3 degrees, secure mind". Preconceptions are my worst enemy.

5) Model Behavior, not Appearance

3D is not 2D plus an extra dimension! Solids in 3D digital modeling just do not sit there to be admired as a static object, like a scripture in a museum. Humans can not perceive 3D unless there is behavior... realistic changes in position, shape, texture, background, lighting, camera, and so on. Modeling behavior is many times more difficult than painting images. In other words, I need to leave my fond memories of PhotoShop far behind. 

6) Need to be Artistic

I next experienced a humbling realization. You actually have to have some artistic skills, in additional to a huge amount of technical skills! Using Maya properly is more than learning the proper sequence of clicks. Maya eventually confronts you and asks, "Are you worthy to use me?" If not, Maya will allow me to rapidly create trash.

7) Modeling Reality is Impossible

My last maturing experience, so far, is the awareness that reality is impossible to model in a virtual space. Regardless of the power of your modeling tool, there will always be some mismatch between your virtual creation and the reality of that creation.

So, you can cheat! Do not model reality but model some fantasy where you can play God. Or, model reality only in a canned production, like a movie, where the audience cannot look behind the curtains or poke the players.

It is a struggle deciding how much realism is sufficient. The art is in the tricks to fudge smartly. As we move into virtual worlds as the dominant media (where we can look behind curtains and poke other avatars), dealing with this issue of sufficient realism will be critical.

Reflections 

My maturing is not done. I have a feeling that Maya can inflict more maturing upon me. And, I believe that my experience with Maya is similar to learning to use properly any complex tool. In my area of Business Intelligence, there are lots of complex tools from vendors like SAP BusinessObjects, IBM Cognos and so on.

There is an industry consensus that pervasive BI is the goal. In other words, everyone in the enterprise should have the skills of using BI tools to improve their job function. In pursue of this goal, we have focused on dumbing-down BI tools to the lowest-common-denominator of capabilities and reserving more powerful capabilities for a few elite power users.

There must be a smarter approach that assists professionals to move gracefully through the maturing stages, as suggested above. If we expect most professionals to use a tool for years to perform their jobs, then it is inhumane for tool vendors to offer either simple tools having limited capabilities or complex tools having great frustration.

The issue of simple versus complex is not relevant. Assisting users to learn the underlying concepts applicable at each stage of maturing is the imperative. Think of it as an Adventure game in which new levels of difficulty are progressively revealed. This is new perspective for designing the help facilities for a professional tool.

At the Boulder BI Brain Trust, we had a discussion with Stephen Brobst of Teradata about the issues and architecture for Operational BI. It is a great discussion! See the complete blog here.

Today I participated in a brainstorming session with 11 persons, most of whom I had not met before. The context was an island in Second Life using voice and text chat.

The session was organized by Erica Driver, a former analyst for Forrester and founder of ThinkBalm, which supplies "independent IT industry analysis and strategic advisory services to technology marketers and Immersive Internet advocates, implementers, and explorers."

We used two simple tools to pool and document our collective wisdom. First, we used a mindmap-like tool to construct a hierarchy of our ideas. View the group with our idea tree. Second, we had a rating device so that participants could quickly vote on an issue. View our rating machine. More images: A and B

Erica will be presenting her thoughts on "Proof Points for Immersive Internet" next Thursday, October 23, at 5:00 PDT. See the Serious Second Life Meetup website for detail. Join us!

There is no doubt in my mind that collaboration using virtual world technology will be mainstream for enterprise BI. It is just a question of when.

Hey, here is a new acronym! You saw it first here! Maybe this blog will be the first and only time anyone will see this acronym!

In the BeyeNETWORK Radio Show for October (mp3 at xx:xx), we had a great discussion of the challenges of using automated decision processes as part of our BI systems. We focused on the stock market, struggled with greed versus fraud, and pondered whether our BI technology could have prevented the instability in global financial markets. My colleagues do not shy away from making earth-shattering statements.

The larger issue is the scalability of BI systems, not in terms of scaling terabytes to petabytes, but in terms of scaling from 10 customers to 10 million customers. In this context, how do we use automated decision processes properly?

The critical factor is how human judgment is embedded within those algorithms. Human judgment is a very limited resource. It depends on the number of persons engaged with the decision process, along with the skills of those persons to understand the complexity of the process, the authority of those persons to effectively influence the flow of the process, and the ethics of those persons to care about the impacts of the process. Quite a challenge!

Let me go out on a limb... Every significant decision process must have some amount of human judgment embedded within it, so as to attribute to a person(s) the responsibility for the impacts of that process. Further, the victims of that process must have recourse back to that person(s).

Yes, wow... Here are two BIG problems.

First, our tools for designing and executing decision process are woefully weak in managing human judgment. We stick in a node that requires a quick YES or NO reply from some poor employee. Please someone prove me wrong on this point.

Second, time constraints limit the extent of human judgment. For example, the space shuttle has thousands of decision processes that must react in milliseconds. The pilot has control, but...

What do you think? Who has done good work on this issue?

Supported by Teradata, Netezza and Kognitio, an independent research study is looking at the trends in the Data Warehouse Appliance (DWA) market. Much is happening, much of which is counter-intuitive.

If you are an IT professional with opinions on DWA, please take 5 minutes to take a survey. I will analyze the results objectively and report those results to you. The complete report will be posted for free on the B-eye-NETWORK. Thank you, Richard

Click Here