Many people who have websites, blogs included, are obsessive about their stats. Always wondering about questions like:

• Which pages are more popular than others?
• Where are visitors to my site coming from?
• How did they find my site?
• How many subscribers do I have to my RSS feed?

It is this last one that has been bothering me at MPOW for a little while now. If you ask this question, the most obvious answer, especially if you run your own blog, is to use a service like FeedBurner.

But what do you do if you can’t use service like this? The blog server I manage at my MPOW will potentially host many blogs. I can’t create accounts at FeedBurner for earch one. It would be too administratively burdensome.

Fortunately kind people like Ken Varnum from RSS4Lib have been thinking about this same question and have done more than think, they’ve done something about it. Ken has built the YourStats tool. You can read more about it on the RSS4Lib blog.

What’s wonderul about the tool is that you can upload your webserver log file, and the tool will work its magic and estimate the number of subscribers you have to your RSS feed.

For those of us, such as me, who can’t transfer logfiles out of our organisations Ken will kindly share the source code of the tool.

Now I’ve got one more statistic in my bag of tricks to help show how popular our blog is.

At MPOW we use a blog to help us communicate with our users. We’ve affectionately called our blog “The ORB” following a successful naming competition that we held earlier in the year.

An integral part of the blog is the Subscribe2 plugin. I think I can safely say that without this plugin our blog would not have gotten past the planning stage. This is primarily because we migrated away from an email list to the blog platform and needed a way to continue this mailing list functionality.

This morning I needed to find out which categories a specific user had subscribed to in order to recieve email updates. The SQL I used is as follows.

SELECT wpt.name
FROM wp_terms wpt
WHERE term_id IN (
    SELECT wpum.meta_value
    FROM wp_users wpu, wp_usermeta wpum
    WHERE wpu.user_login = 'USER'
    AND wpu.id = wpum.user_id
    AND wpum.meta_key LIKE 's2_cat%' )
ORDER BY wpt.name

Replace USER with the code above with the user name of the user you are investigating

Yesterday I needed to get a count of the number of posts in a WordPress blog in each category. I used the following SQL code.

SELECT wpt.name, COUNT(wptr.object_id) as post_count
FROM wp_terms wpt, wp_term_taxonomy wptt, wp_term_relationships wptr
WHERE wpt.term_id = wptt.term_id
AND wptt.taxonomy = 'category'
AND wptt.term_taxonomy_id = wptr.term_taxonomy_id
GROUP BY wpt.name

The same technique will work with WordPress MU based blogs as well. Just substitute the table names with those for the specific blog in the WordPress MU system.

Something to keep in mind is that the if you add up the total of the ‘post_count’ fields reported by this query it is likely to be more than the total number of posts in your blog. This is because posts can be in one or more categories.

I captured this tunnel of water drops as the water came out of the spouts in the second of the two fountains in front of the State Library in Adelaide, South Australia.

(Click image for larger version)

Earlier today I needed to copy a non encrypted Digital Video Disc (DVD) to make a few backup copies. The disc contained a movie recorded on a video camera of someone that MPOW is collaborating with on a project. As always I used my laptop running the Ubuntu Linux distribution.

I first tried to use the Brasero Disc Burning application to make the copy. Unfortunately the application tried to make the copy as if it was a DVD containing files and data and not a DVD containing a movie. This meant that the copy was corrupt and would not play.

To make a copy of the disc I made a copy of it first as an ISO file by using this command:

cat /dev/dvdrw > ~/backup-copy.iso

Please note that the /dev/dvdrw path may be different on your system, for example it may be /dev/cdrom. The filename, after the >, is also arbitrary.

Once the copy using cat completed, which took some time, I was able to use the Brasero Disc Burning application to burn the ISO image just like any other image.

This technique resulted in a perfect copy and I can use the ISO image file as many times as I need to make as many copies as I need.

The “Water droplet on a DVD” photo was uploaded to Flickr by -j0n-.

Last night I decided to set myself a challenge, possibly for the beginning of a new project. I wanted to install the latest DSpace open source repository software, currently version 1.5, on an Ubuntu based environment.

Knowing that I’d need to have things like a PostgreSQL database, and have Apache Tomcat running, I wanted to explore this project in a virtual machine environment. I wanted to do this so as to reduce the likelihood of it interfering with the existing services and development environment I have on my laptop running Ubuntu.

Below are notes about the things that I learnt. This isn’t in any way meant to be a full procedure manual, it is more about highlighting potential issues that I experienced along the way and hopefully act as some documentation for myself at a future time.

Virtual Environment

In the past I’ve used the Virtual Server software from VMware for virtual machines. The last time I used virtual machines, around 6 months ago, I had issues with time synchronisation and the compilation of kernel modules. Apparently at the time the Linux kernel I had running on my Ubuntu installation was newer than the one supported by Virtual Server.

I looked around for an alternative and found VirtualBox which is software from Sun Microsystems. I was pleased to see that the latest release official supported Ubuntu 8.04 which is what I’m running. Sadly the latest release isn’t in the Ubuntu repositories yet. Fortunately the deb installation file from Sun worked really well and the installation was very smooth and is running perfectly for me.

Server Operating System

Being most familiar with the Ubuntu distribution of Linux I decided to use the Ubuntu Server edition for my virtual machine. In this way I wouldn’t be encumbered with a graphical user interface. My virtual machine could be “lean and mean”.

One thing to note is that to use Ubuntu Server edition inside a VirtualBox virtual machine you must enable the PAE/NX option so that this feature of your CPU will be exposed, and therefore used, by your virtual machine. This is an option that you can set on the “Advanced” tab on the “General” settings display.

Additional Software

For the majority of the additional software required by DSpace I used the versions in the Ubuntu repositories. This includes things like the PostgreSQL database, the Sun Java JDK, the Apache Maven build system, and the Apache Ant build system.

I used the Apache Tomcat 5.5 software from the official Apache Tomcat website. I did this so that I can run both it and DSpace under its own user dspace user. This makes life so much easier. Trying to the Ubuntu package of Tomcat was not working. I kept getting errors related to permissions.

DSpace installation

These are brief notes about the installation itself:

• The documentation is not kidding when it says to use a specific dspace user.
• I’ve installed everything DSpace related into the home directory of the dspace user.
• The build time will take about 15 minutes, based on my experience. Most of this time is taken up with the Maven build process downloading a seemingly endless list of modules.
• Once build the installation is very straight forward.
• Installing the options Tomcat administration package can be very useful in debugging things.

Final Notes

Once installed and running the latest version of DSpace looks just like it did the last time I installed it. This is deceptive as their are many new things in this new version, all very interesting things that I want to explore.

The latest version cements in my mind the opinion of DSpace I formed while I was on the RUBRIC project, which seems like a life time ago now. DSpace is a solid option for those wanting to implement a repository and shows that you don’t need a vendor supported solution, as some other projects and vendors would like you to believe.

I look forward to exploring some of these new options in the near future.

I captured these water drops coming out of one of the spouts of one of the two fountains in front of the State Library in Adelaide, South Australia.

(Click image for larger version)

“To get something done, a committee should consist of no more than three men, two of whom are absent.”

- Robert Copeland

Spotted on the The Lone Sysadmin blog.

A few weekends ago I was walking around the Adelaide CBD and stumbled across the Pioneer Women’s Memorial Gardens. In it is this statue.

(Click image for larger version)

I’ve been seeing one of my favourite photographers taking a number of B&W images and so I decided to try it with this statue. They came out rather well I think. The shadows and highlights created by the use of B&W add to the image I think. This is the statue from another angle.

(Click image for larger version)

An inscription at the base of the statue reads:

“The garden of remembrance and the flying doctor base at Alice Springs  were established in the centenary year 1936 by the women of South Australia as a tribute to the pioneer women of the state”

In a recent programming task I’ve been working on I’ve needed to transfer some data, in the form of arrays, from one PHP page to another via hidden input fields in a form. I’ve used JSON for this task. More information on JSON is available at the official website.

Some of things I’ve learnt are:

Check to ensure the JSON functions are available

The JSON functions in PHP are provided by an extension that is only bundled with PHP version 5.2 and above. If you don’t have that specific version you can compile and include the php-json extension yourself, or use the JSON-PHP library which is written in pure PHP and doesn’t require any compilation.

Detecting which library to use is easy

To ensure your code doesn’t cause problems you can construct it in such a way that it detects if the php-json extension is available and if not use the json-php library using code like this:

if(!function_exists('json_encode')) {
    // PHP pre 5.2 or the library isn't installed
    // include the alternate library
    require_once('path/to/library/JSON.php');
    $json_services = new Services_JSON();
    $encoded = $json_services->encode($my_var);

} else {
    $encoded = json_encode($my_var);
}

The same technique will work for decoding as well:

if(!function_exists('json_encode')) {
    // PHP pre 5.2 or the library isn't installed
    // include the alternate library
    require_once('path/to/library/JSON.php');
    $json_services = new Services_JSON();
    $my_var = $json_services->decode($encoded);

} else {
    $my_var = json_decode($encoded);
}

Including the data in a form requires some extra work

To include the data in the form, you need to change the double quotes (”) in the encoded data into single quotes (’). Otherwise the generated HTML will be invalid. You can do this quite easily using the str_replace function. Remember to do the reverse replacement before decoding the data.

Ensure you get the variable type you expect

By default an array will be decoded as an object. This can make using it in your code a little bit more complicated. Fortunately the json_decode function will take an additional boolean operator to return an array, and you can use variable type casting when using the json-php library. For example:

if(!function_exists('json_encode')) {
    // PHP pre 5.2 or the library isn't installed
    // include the alternate library
    require_once('path/to/library/JSON.php');
    $json_services = new Services_JSON();
    $my_array = (array) $json_services->decode($encoded);

} else {
    $my_array = json_decode($encoded, TRUE);
}