clariovoyance

We strongly believe two keys to success for the clario platform will be transparency and engagement. Our experience is that the ideal clario user is one who sees clearly the direction clario is headed and has the opportunity to help us define our many meanderings along the way. With this vision in mind, I want to share our current development priorities with you. Note: "current" is the key word here. I'll be posting the roadmap on clariopedia so you can watch it morph over time.

I'll dig more deeply into each area in future posts, but for now a high-level summary:

• Standard Nodes: Research and experience guides our activities here. Bivariate entered beta in clario 1.3.0 this week, clustering (K-means and EM) development is underway with decision trees to follow.
  
• Continuous Improvement - Usability & Performance: This is natural progression, driven heavily by user feedback. We are constantly asking ourselves how we can make clario more intuitive and better performing (in both the user interface and data crunching). Aggregate, Missing Values and Join are all queued for node configuration redesigns.
  
• Productivity & Usability: These are significant enhancements (versus the more incremental changes above) to usability and the user interaction model. The primary focus of the team's activities for clario 1.4.0 is the implementation of "Non-Destructive Metadata". clario currently uses a destructive metadata model where upstream changes to node configuration results in a reset of all downstream nodes. This was a developmental necessity (simply from a time-to-market perspective), but one that causes our users tremendous frustration. Implementing graceful "conflict resolution" for these upstream configuration changes, which minimizes user rework, is our top development priority.
• clario Analytic Web Services Enhancements: Oddly enough, the clario platform can be leveraged without ever seeing the user interface (I know...shameful). Our goal is to allow organizations of all sizes to leverage the power of the clario platform seamlessly from their own applications via web services.
  
• Web Services Nodes: Perhaps our initial proof of concept explains it best: the USPS node will allow users to feed a file containing customer address information through the USPS Address Standardization process seamlessly within a single clario workflow. Other web services nodes will allow you to pull data in directly from third-party systems to create a comprehensive customer view.
  
• Massive Volume Scalability and Speed: clario has two limitations today, no single persistent file can exceed 5 Gb (compressed) and no file with more than 1,000 attributes can be read in. A long-term initiative is underway to leverage the power of Hadoop and the MapReduce model to drastically enhance the scalability and speed of the clario platform.

I hope this overview was helpful. Most importantly, I hope each and every one of you will leverage the forums on clarioanalytics.com and the support email address to submit your comments, questions, suggestions and concerns.

Matt
Co-Founder & CTO
clario Analytics