Some sort of disclaimer

• Added support for aspects using the @Aspect annotation. Checkout the demo for some cool examples.
• Clear distinction between folders and content nodes using two separate base classes: ContentNode and FolderNode
• Several other bugfixes

As in most common designs, you usually separate these services in a dedicated layer that aggregates and uses core components like NodeService, ContentService, etc. However, since the dawn of time, Alfresco supported only a “generic” approach for content manipulation, mostly achieved through calls to get/setProperty() on an injected NodeService bean:

    String name = (String) nodeService.getProperty(nodeRef, ContentModel.PROP_NAME);

Pojonode is a library designed to move away from this non-OOP programming model and provide a POJO-based alternative for it. The same way that JPA and other ORM technologies enable the mapping between a pojo-based model and a relational database, Pojonode enables the mapping betwee an object-oriented model and an Alfresco content model definition. Pojonode’s techniques are somehow similar with the classic ORM approaches, but with some important extensions/adaptations to Alfresco’s core concepts.

You can read more about the framework on the Google Code project page which is a good starting point for understanding Pojonode’s main concepts and purpose.

While there are certain features still in work (like wrapping ContentService calls), most of the core Alfresco content management concepts (like managing properties and associations) are currently implemented and working.

Other features that might come in handy are the generation of content model XML and web client configuration. When enabled, they can handle the complete setup of the Alfresco content model for you. Check the Integration section of the wiki for details.

For whatever reason, the gnome-settings-daemon in the Gnome version shipped with Ubuntu 10.10 eats all your RAM. By all I literally mean ALL. I have 4GB, and in a few hours it was using 3GB. Pure awesomeness.

Is wonderful how completely useless background processes end up doing this kind of crap. Why doesn’t this ever happen to MySQL or JVM? Motherfuckers…

* http://ubuntu-virginia.ubuntuforums.org/showthread.php?p=10051835
* https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/485923

Some sort of disclaimer

They don’t provide any public lightweight API, from what I know, so a lot of HTML parsing is involved in getting this application working. A bit slow, of course, but hopefully this will change if BOG decides in the future to expose their service through a JSON API or something.

The application is not yet on the Android Market, but hopefully I can get it up there once I clean it up a bit. For the moment you can get it from here.

Some screenshots of the application running on an HTC Desire:

I used the Nikkor 135mm f/2.8 AI-s for most of my pictures, which is all manual on my D90 (yeah, including focus). But I managed to get a few decent shots: http://www.flickr.com/photos/cosminmarginean/tags/peninsula2010/.

Some sort of disclaimer

Anyways, going through the list, the candidates either look over designed (and clearly serving other purposes), were licensed under some inconvenient terms or just with a low level of control over the features and content aggregation.

So I decided to give it a shot and see how really complicated is to code something like this with flex and here are the results: http://cosminaru.ro/photo/cosmin and http://cosminaru.ro/photo/alexandra.

It doesn’t have the smooth scroll and other fancy stuff but it gets the flickr sets and it displays their content – which is actually what I was looking for. Later maybe I can find some time to add some cool effects here and there, but for now is good enough. It also has support for linking directly to one of the images so you can do stuff like this.

The binary package can be found here. It works by editing the flickr-user.txt file with your flickr user id (not user name) which you can obtain from here: http://idgettr.com/.

I also want to publish the code but it needs some cleanup and some wtf’s removal first.

The original Excel file published on a web server:

Downloaded this on Windows to fuck around with it.

Download on Ubuntu to edit and etc…

Some sort of disclaimer

Some sort of disclaimer