Add comment September 12th, 2010
Running java applets requires the browser to recognize the java installation on the computer. OK, that was a no brainer. I am now running Firefox 3.6 on Ubuntu 10.04 LTS with Sun’s JDK 6 installed and a app in a website failed to run. It displayed an annoying message that Java was not installed. But like I said, I did have java installed. I am a java developer and have been doing java things all along. So naturally, I went huh!. I looked into all Firefox options and could not find anything that I could get Firefox to point at my java installation. Finally I stumbled upon the below firefox plugin after some google searches. Installed that and all was good. A good webpage to test java in a browser is Javatester.
IcedTea NPR Web browser plugin
Although, java in my browser was working, it changed my /usr/bin/java link to OpenJDK’s installation that came with this plugin. If that happens to you, then run the below command and reset your JDK appropriately.
sudo update-alternatives –config java
Add comment August 18th, 2010
I installed the latest Ubuntu 10.01 LTS Desktop. To my surprise, I found Ubuntu has dropped Sun’s JDK from its repositories and instead recommend OpenJDK. But I really wanted to use Sun JDK. For no other reason other than, I have been using it and did not have to time and patience to test Open JDK and may be deal with whatever was different.
Wierd thing is when browsing to http://java.com, the instructions are arcane. So here is how I did it.
sudo add-apt-repository “deb http://archive.canonical.com/ lucid partner”
sudo apt-get update
sudo apt-get sun-java6-jdk
Now back to work. Cheers.
Add comment August 16th, 2010
I just read this blog Why Sentiment Analysis Can’t Work And Why It’s A Damned Good Thing. The examples in there are very good at explaining the problem.
Add comment July 19th, 2010
An ubuntu tip:
How to check the ubuntu version installed
$ sudo lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 8.10
1 comment November 24th, 2009
If you are running a 64 bit OS, how would you know if the installed MySQL is running on 32 bit binaries or 64 bit? There may be several ways to do it. Here is one way.
$ which mysqld | xargs file
Response when running 32 bit MySQL is as below.
/usr/sbin/mysqld: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.6.8, dynamically linked (uses shared libs), stripped
Response when running 64 bit MySQL is as below.
/usr/sbin/mysqld: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), for GNU/Linux 2.6.8, dynamically linked (uses shared libs), stripped
1 comment September 17th, 2009
By default any 32 bit OS should see less than 4GB of RAM available to it. Your motherboard can support upto 4GB but generally you will see something like 3.1GB to 3.5GB showing up for use by the OS. It has to do with the virtual address space that a 32 bit OS can address (2^32). But I wanted to use 8GB of RAM. So I googled and found a solution that worked for me with Ubuntu 8.10.
Check BIOS capability
$ dmesg | head -25
You are looking for a BIOS memory mapping to indicate that the section above 0000000100000000 is usable. If you don’t see that, see if you can upgrade/update your BIOS and enable “memory hole remapping” or “memory hoisting” in your BIOS .
[ 0.000000] BIOS-e820: 0000000100000000 – 0000000134000000 (usable)
Check if your CPU is 64 bit
$ grep flags /proc/cpuinfo
If you see lm, then it is 64 bit. (you will lm in red color, don’t know why)
Then install linux server version kernel
$ sudo sudo apt-get install linux-headers-server
$ sudo apt-get install linux-image-server
Add comment April 7th, 2009
yum install php-mysql mysql mysql-server
/sbin/chkconfig –levels 235 mysqld on
Add comment February 26th, 2009
There are efforts underway for building platforms that will enable users to ask complex questions in plain English and be able to get back useful answers for analysis. Most efforts focus on labelling text data (unstructured data) with semantic metadata. This works well for web data which is almost exclusively unstructured data ( web pages, blogs, wikis, feeds, Discussion boards, etc. ). Enterprises though have years or decades of valuable information locked up in databases/data warehouses. In the last decade several BI vendors have made strides in unlocking the value hidden in there. Many enterprises have worked on exposing this data as web services for consumption in various applications. But this data lacks the semantic annotations that would enable it to become alive for free flowing user language queries.
With the current trend of providing RESTful web services to allow access of the web data, GoogleData, Freebase, Technorati, and infinite others are making the web look more like database platforms. On the flip side, if datamodels in the databases could be annotated with the conceptual models of knowledge space provided by Ontologies, the databases will become semantic web service endpoints sporting SPARQL protocol. Then this merging of the Web and Enterprise should bring in a new wave of BI innovations.
Add comment October 31st, 2007
Semantics is data describing content in a structure that is machine readable. Freebase (like GoogleBase) is a graph structure that is allowing collective intelligence of the web help it build a sematic network of topics. It calls itself a open database of the world’s information. They have a api that is completely free and has limitless reads for now. Freebase schemas are gettting mature very quickly. Freebase can be very successful in providing the semantic structure for topics where Wikipedia could not.
The company schema in Freebase should have links to competitor companies. The type of the link could be further described to be a sector level, product level, company level competitor and so on. In the same line of thought, it might be extended to customers, partners and so on.
Hoovers, D&B, InfoUSA are putting a lot of human effort in compiling competitive intelligence are market data. Freebase though has the potential of taking off like Wikipedia and could easily cover the data carried by these vendors. ZoomInfo has been using machine intelligence and collecting web data. But they have serious quality issues when it comes to the long tail. This is an even bigger issues with D&B (ZapData). For Freebase though, as the data is open for applications via API’s the quality of data will also eventually get stable.
Add comment October 31st, 2007