Published by Stef May 16th, 2008
in Technology.
Some folks asked if we watch too much TV here at Netrics… why did we name our blog “Netrics HD”? Are we launching a high-def channel? Alas, no we aren’t.
We chose the name because we believe in viewing database data – regardless of the application – though a “High Definition” lens. It’s this High Definition lens that enables us to deliver High Fidelity data, despite the fact that the data itself is not perfect (moreover it can never be perfect).
This High Definition lens happens to be powerful mathematics that learns about data. As opposed to rules-based (probabilistic and deterministic) solutions that require a lot of guessing.
Published by Stef April 18th, 2008
in Technology and Business.
Recently a friend of mine asked exactly how I was planning to blog about both business and technology. He pointed out the well known (but flawed) theory that says business folks are bored with technical details and technical folks are bored with business details.I love both business and technology and am particularly fond of their intersection!
I hope to offer meaningful insight on both subjects and, of course, how they need each other.
Published by Stef April 16th, 2008
in It Happened.
All these butcherings of my last name remind me of something that happened a few years ago.
We were meeting with a large hospital near the west coast and, as I often do, I gave a few examples of my own experiences having “Damianakis” butchered (how can you not butcher such a name?) and how that affected my experience with various organizations. And of course, the flip side, how the organizations were impacted as well (added expense and risk, etc).
At one point during the meeting I was fielding questions about the advantages of our mathematical modeling approach when the CIO raised his hand and asked me “is Damianakis your stage name?” I was caught off guard, “No, I’m not involved in the performing arts” I responded. “No”, he said, “is that your real name?” Then it finally registered…
It’s good to know that I have the right name for my chosen profession.
Surprisingly though, most people suffer the same problem with their names as well. At least that’s the result of my informal, unscientific poll. Even people with short common names often end up having their names being butchered. Common responses in my poll were “they never get my name right” and “they often make a mistake,” etc…
Unfortunately computers are equal opportunity offenders, so the problem isn’t confined to just names of people, it happens with every kind of data they store.
Published by Stef April 15th, 2008
in Technology and Business.
Alas, this core matching problem is not limited to people names… it basically applies to anything and everything that is stored in your DBMS. For example, product names, vendor names, claims, applications, and the list goes on and on. No data is immune.
This fundamental data matching limitation then spreads to every application built on top of a DBMS.
So this seemingly small functional limitation – the inability to intelligently compare and match imperfect data – can, at best, render enterprise applications ineffective and wrong, or worse still, result in their failure to perform as required.
Very scary… but fear not, all is not lost! More details on this is on the way… mixed in with other commentary of course.
Published by Stef April 10th, 2008
in Technology and Business.
We often get asked what creates the business-data problems that Netrics solves… Why is database data not perfect? How can databases make mistakes?
These business problems arise because of computers’ well-defined limitations - after all, they’re just fancy calculators. Calculators that can perform billions of operations every second - quite amazing when you think about it. Built into every CPU is the capability to compare data - but these data comparisons are limited to “equals” or “not equals” - that’s all.
These same limitations propagate up to the applications (i.e. databases) that run on these CPUs. So, if you ask any commercially available DBMS (e.g., Oracle, IBM DB2, MS SQLServer, MySQL, Cache, etc) to retrieve a record where the last name is “Damianakis” then that’s exactly what it’s going to do.
But what happens if “Damianakis” is actually in the database as “Damanakos” and also as “Dananakis? Easy, the DBMS will find neither - at least not with built-in database operators. Someone needs to build an abstraction layer that uses the built-in database operators to address this issue.
And thus the can of worms is opened…
Published by Stef April 8th, 2008
in Technology and Business.
Finally! I’m excited about the launch of our new blog. So, what’s this blog all about? Well, here at Netrics we’re passionate about the business and technology of data matching. An esoteric topic you might think, but as we’ll explore over the coming blogs-to-be, a technology that’s mission-critical for virtually every IT solution deployment.
The goal of this blog is to educate and evangelize while offering a “high-definition” view of data matching. This means we won’t sweep details under the rug, or obscure them with marketing-speak – we feel it’s important to see and understand how data matching works and the broad impact is has on IT solutions that businesses depend on.
I’ll certainly get ideas, suggestions and help from everyone here at Netrics… Here is a partial list of topics that I’ll cover in upcoming posts:
- how does inferior matching performance adversely affect a company’s bottom line?
- real-world examples of organizations’ failures to make effective use of structured data due to matching issues, and the impact of these failures on the ability to conduct business
- how ineffective data matching hampers organizational compliance efforts, resulting in increased regulatory issues with government agencies
- why is real-time error-tolerant matching computationally expensive?
- what is deterministic matching? what is probabilistic matching? how are they different?
And of course, I’ll look to you, the readers, for comments, feedback and topics of interest. I’m very excited to have a medium to talk about these mission-critical business problems and the innovative mathematics that Netrics developed to solve them.
Recent Comments