Cold War all over again?  Not really - it’s obvious what happened here.  But I think this is a great example of how the semantic web will wrap people, places, and things with context, and these types of issues will become a thing of the past!

Instead of “Georgia”, a semantic agent will see something like http://dbpedia.org/page/Georgia_(country) or http://dbpedia.org/page/Georgia_(U.S._state) and thus have an understanding of how to differentiate the two.

The problem I ran into was with the avatar upload.  When I tried to upload my avatar, I would be greeted with a blank page (the “oh-crap-internal-server-error-where’s-my-whiskey” kind of blank page).  I opened up my /var/log/httpd/error_log to seek out what exactly occurred.  Here’s what I found:
[Wed Jul 23 13:01:55 2008] [error] [client 10.6.132.139] PHP Fatal error: Call to undefined function image_type_to_extension() in /var/www/html/laconica/classes/Profile.php on line 85, ...

Interesting.  My brand new fancy avatar is now reduced to a broken image in Laconica.  To revert back to the default gray and apathetic face avatar, simply remove all entries in the avatar table of your Laconica DB instance.

Some quick background about the configuration:

$ uname -srvpio
Linux 2.6.18-92.1.6.el5PAE #1 SMP Fri Jun 20 02:51:01 EDT 2008 i686 i386 GNU/Linux

$ php -version
PHP 5.1.6 (cli) (built: Jun 12 2008 05:02:35)
Copyright (c) 1997-2006 The PHP Group
Zend Engine v2.1.0, Copyright (c) 1998-2006 Zend Technologies

A fairly recent setup.  Since this was a fresh install, I figured something had to be wrong with the installation of the PHP GD libraries.  Naturally the next step would be to do a var_dump(gd_info()); to confirm any issues:
array(12) {
["GD Version"]=>string(27) "bundled (2.0.28 compatible)"
["FreeType Support"]=>bool(true)
["FreeType Linkage"]=>string(13) "with freetype"
["T1Lib Support"]=>bool(false)
["GIF Read Support"]=>bool(true)
["GIF Create Support"]=>bool(true)
["JPG Support"]=>bool(true)
["PNG Support"]=>bool(true)
["WBMP Support"]=>bool(true)
["XPM Support"]=>bool(false)
["XBM Support"]=>bool(true)
["JIS-mapped Japanese Font Support"]=>bool(false)
}
Nope. Everything looks fine with the setup.  And according to PHP.net, image_type_to_extension() should be available for >=PHP5.0.  This, however, doesn’t appear to be the case.  As Chris notes:

Function was added in PHP 5.2, not PHP 5 (you would assume this means 5.0)

So, evidently, my current version of PHP (and the most recent in the RHEL5 repos) does in fact not support the image_type_to_extension() function. If you have the most recent version of PHP, you should be fine.  If not, here’s the fix: To work around this issue, I devised a simple modification that would overwrite the image_type_to_extension() functionality if you’re experiencing the same issues. Head to $LACONICA_HOME/classes/, back up Profile.php, and then open it up in your favorite editor. Head to Lines #84 and #85 $filename = common_avatar_filename($this->id, image_type_to_extension($info[2]), NULL, common_timestamp()); and replace it with the following code snippet below:

# Start - More RAM.
if(function_exists(image_type_to_extension))
{
$extension = strtolower(substr(image_type_to_extension($info[2]),1));
}
else
{
list(,,$imageType) = getimagesize($image);
switch($imageType)
{
case IMG_GIF:
$extension = ".gif";
break;
case IMG_JPG:
$extension = ".jpg";
break;
case IMG_PNG:
$extension = ".png";
break;
}
}
$filename = common_avatar_filename($this->id,$extension,NULL, common_timestamp());
#END - More whiskey.

Save Profile.php, and now try to upload a new avatar. Problem fixed!

The areas of technology that interest me, as you could probably tell, are within the areas of portable data (DataPortability) and semantic web (Linking Open Data).  I see these as two complimentary ways of diminishing the effects of “information overload” (and there’s certainly some overlap here):

1.  DataPortability aims to reduce the instances where data is redundantly generated (such as social network contacts between MySpace and Facebook) by promoting the implementation of open data format exchanges between heterogeneous web services.

2.  Semantic web technologies aim at redefining how information is located and retrieved.  To be able to quickly and precisely acquire information means not having to deal with as much “noise” surrounding search results, and hence, information overload is abstracted away from the user.  (Next post is around some insight around this where Sherman Monroe at Linked Data Planet likened conventional Internet search to finding detailed photographs on a low resolution monitor.)

I think these two related initiatives aim to cure information overload from a technical standpoint.  But what are some other initiatives that aim to bring some sanity to the deluge of information that bombards our screens every day?

Here’s an interesting organization that I was just informed of from a work colleague called the Information Overload Research Organization (IORG).  According to their site, their mission is to:

“…work together to build awareness of the world’s greatest challenge to productivity, conduct research, help define best practices, contribute to the creation of solutions, share information and resources, offer guidance and facilitation, and help make the business case for fighting information overload.”

The organization is planning a day-long “inaugural” event filled with breakout sessions to discuss topics surrounding information overload.  Although I’m not entirely certain the exact nature of what will be covered, I was interested to see some of the IORG members on the list - including Microsoft and Google.  I noticed a few members who work in the areas of Human-Computer Interaction (HCI), which got me thinking a bit. We’re tackling information overload from a data perspective, but will this ultimately solve all of our problems?  How would I render and display a complex query again semantic metadata that returns back thousands of triples?  This is where I see portable/linked data spanning into other disciplines such as HCI.  Can conventional web interfaces handle all of this robust data so that it’s intuitive to the user?  There’s certainly a lot more room for different concentrations to attack these issues.

I threw up the IORG Inaugural Event in Yahoo! Upcoming here, as I certainly plan on attending with my schedule permitting.  I think a lot of the things discussed there will in fact complement the more technical focus around portable/linked data.  What do you thing?

Is anyone more familiar with IORG and would like to comment?  I’m certainly interested in hearing more about the organization and the event!

This blog is now “OpenCalais-enabled”.  They have a Wordpress plugin called Tagaroo which reads your Wordpress blog entry as you type it, and automatically suggests tags for it.  It also pulls down Flickr images based on these tags so they can be included with the entry.  More in-depth review of these tools after the presentation!

But here’s something that people may find interesting. For my INFO633 (Information Visualization) class at Drexel, I developed a framework for tracking trends and patterns in Google Earth based on concepts captured from APML profiles. This is my initial research in the feasibility of the idea, spanning about four weeks of work. I submitted it to the IEEE Information Visualization 2008 conference for review out of interest for understanding the call for papers process.

I’m a huge fan of DataPortability and am actively seeking areas where regular (non-technical) users will find value in the concepts and principles outlined by the initiative. Possibly a project for the DataPortability Labs?

Download the file here (11.3MB PDF warning).

Development is on-going. Comments welcome!

APMLStream uses the XStream and XML Push Pull Parser libraries, and is currently released under the Apache License 2.0.

You can browse the source here at Google Code, or checkout and build the source from SVN:  svn checkout http://apml-library.googlecode.com/svn/trunk/Java/APMLStream/APMLStream-0.1/ apmlstream-src

To build the source, you’ll need the Java 6 SDK and Apache Ant.  I’m using the most recent Java SDK (1.6.0_04) and Apache Ant (1.7.0).  Building is as simple as issuing the following commands in your terminal:  ant, ant jar, ant javadocs.  Then, include the apmlstream-0.1.jar into your project, and you’re done!  I’ve also provided the binary distribution with this post.

Although I still have a laundry list of to-do’s for this library, v0.1 contains the basic functionality for APML data manipulation. Expect an APMLStream tutorial within the next day or so! Comments and suggestions are also welcome.

Downloads:apmlstream-0.1.jar apmlstream0.1-javadocs.zip

My IT street-cred was at stake…and I couldn’t idly sit by and watch it wither away.

I have loads of neat stuff going on in my life, that I’m anxious to post here. Please bear with me as I finalize the page design. Thanks!