Ok, this will be short. XMLA, for those who have never heard of it, is XML for Analysis, a standard for connecting OLAP service providers to front ends like query builders, OLAP browsers etc. Why do we need it? Because monolithic solutions in OLAP domain raises the cost and decreases the flexibility of solutions, but interoperability in this domain seems to be going down. The XMLA council seems to be inactive for a long period of time, and XMLA home page is taken over by Simba technologies. ...

There was this TV show which I used to love. I was fascinated actually. It was about a family who has accidentally passed onto some alternative reality. They were trying to get back, while escaping from some guy , and they were trying to fit into weird cities they were travelling. It is one of those things that is carved into your mind when you were a little kid. I do not know why, but I felt an urge to find what it was actually called, for it was aired when I was a little kid, and I found it. Check it out here in case you want to know about a series called otherworld. I’ve always been a fan of science fiction, and this one was fun. Too bad it is not possible to have a dvd set of the series.

I’ve been working on implementing a data warehouse based on open source tools. The actual information system that produces the data runs on Microsoft Sql Server and other MS technologies. I have been asked to implement a minimum cost solution, zero if possible. It appears that some parts of the open source business intelligence domain are more mature than others, and zero cost might not be possible after all. KETTLE project from pentaho looks good enough for ETL. I’m using Mondrian (again from Pentaho) as an OLAP server, and even though it does not support complete mdx implementation, it looks promising for the moment. However, the front end is not promising. JRubik is no longer developed, and JPivot is in need of a face lift. Recent trends in web development raised the bar for UI requirements, and without even talking to the clients I can see that current solutions won’t cut it. ...

Ok, let’s just accept it, web is not what it used to be anymore. I will not go into details but web user interfaces have accepted and embraced javascript. About five years ago, I was busy trying to build a user interface based on javascript tricks for an electronic claim processing system, and it was a nightmare. Firefox was not a major consideration back then, but internet explorer had threee version that ran on 3 different version of windows. I have suffered a lot from inconsistent apis, and I had to invent a lot of tricks similar to basics of ajax today. ...

Ok, I’ll be honest, I’ve always been into probabilistic methods, for they somehow “fit” into my way of thinking. There is something about probabilistic methods, and probability theory; you are either suitable to work with it or not; you either love the field, or hate it. I’m the kind of guy who has some love hate relationship with it. I certainly like the field, but the overall concept is so deep and abstract that I can get lost very easily. Something that makes perfect sense seems like Chinese the next day, but I still can’t let go. ...

Ok, I’ll be honest, I’ve always been into probabilistic methods, for they somehow “fit” into my way of thinking. There is something about probabilistic methods, and probability theory; you are either suitable to work with it or not; you either love the field, or hate it. I’m the kind of guy who has some love hate relationship with it. I certainly like the field, but the overall concept is so deep and abstract that I can get lost very easily. Something that makes perfect sense seems like Chinese the next day, but I still can’t let go. ...

I have a project in which I have to come up with a basic data warehouse implementation. I have to deal with all the basics, ETL, Cube design, etc, and on top of that I intend to build a naive Bayesian classifier generator for decision support. (I might consider ID3 or C 4.5, but I’m not sure they are free for this kind of use). Developing all of these from scratch is out of the question, after all why should I do it if I do not have to. Having a decent UI at least for some tasks would be nice though, and Pentaho might be the answer. I have been following Pentaho for quite some time now, and finally I need exactly what they provide, for a consultancy job. I guess we’ll see if they are up to the claims they make. Most of the parts of their product portfolio are based on well known tools like weka or mondrian, but they have been building solutions that use eclipse rcp to wrap these tools, and might be able to do a lot with their existing solutions. I’ll write a detailed summary of my experience, but for the moment Pentaho seems to be the only vendor that opens a free, open source solution. If I can reuse their work, that’d be a really very important base for my future plans, because I’ve always believed that business intelligence and/or analysis tools require knowledge in various areas like data mining, machine learning etc, to provide a real benefit. So money paid for any of these tools should actually be paid for the expert not the tool, since I can hardly imagine an off the shelf tools providing the real benefit of the mentioned concepts. Well, I guess we’ll see about that.

In the past, scripting has been a real strength of Microsoft technologies. I’ve seen some very capable solutions built with visual basic, and as new generations of technologies emerged, I expected Microsoft to provide more robust solutions. Scripting is important, because it lets you come up with flexible solutions. Most of the arguments which are against scripting for large scale development focus on weak type checking, and performance. However, scripting has one very important advantage: when you build a core set of functionality which covers the basic requirements of a domain, scripting gives you the ability to provide this functionality to others who might have less information about the domain. ...

Ok, this is a rant, so be prepared to hear some whining. I have spent almost two days trying to compile Apache Axis2, for I needed a java web services solution. I could have used the latest binaries, but the problem is there is a bug in the eclipse plugin source, and IF I can get my hands on the plugin sources, I can fix it. I won’t get into story of my frustration, but let me tell you this. At the moment, the latest downloads of Apache Axis2 can not be compiled using the instructions in the bundle. I have failed to find other instructions, and I had to switch to Metro. It is really interesting that such a well known project is not providing a source bundle which you can compile by following instructions. Some of the subdirectories (like the eclipse plugin dir) contains instructions which use maven1, and the main readme uses maven2. ...

Well, it appears Google and Microsoft are going to clash once more. As you can read here Google is going for PHR too. Not many details for now, but I’d like to see how these two 800 pound gorillas are going to tackle the usual problems. Healthcare IT is tough, it is so tough that even these two giants can get burned. Let’s wait and see.