Announcing IRM 2, MkIII

I’ve done a lot of thinking, some non-thinking, talking to people, users, and some (well ok, a lot) of prototyping. As a result of all that work, I have decided to revive and continue to work on a new IRM 2, the re-architectured Information Resource Manager application. New features, new code base, but same target, you, the dedicated IRM users.

I’ll get into all the details of what IRM 2 is going to do for you, but I’d first like to share some history with you about where IRM has come from, the previous short attempts at IRM 2, and how its influenced my current direction.

In the beginning was the command line.

Then came IRM. It was quite literally the second piece of software I wrote as a web application. It used the trendy new (well at the time) PHP 3.0. MySQL 3.22 was the in thing. The whole LAMP architecture was just in its infancy.IRM was created way back when (1999 for those keeping score) for my high school. It allowed people to post problems (using the odd nomenclature “Tracking”) about computers that were in its database. The database seed script that comes with IRM reflects these early beginnings, including the computer named Ants and the Library Back Room.IRM went through several versions, 0.5, 0.9, 0.9b, 0.9c, etc. Around the 1.0 release, a really great guy from the University of Tulsa , Keith Schoenefeld started contributing to IRM. He cleaned up a ton of my early major misguided attempts at software development (turns out ENDIF; is kinda not cool), and over time provided new functionality such as the FAQ and Knowledgebase system.

IRM grew in leaps and bounds, adding the Software subsystem, the Reports subsystem, and then Networking. Patch submissions from Mica Curie also started, and some key features were added to the project.
In 2004-2005 another contributer joined the project, Matt Palmer. Not only did he make a nice Debian package for IRM (apt-get install irm is so much easier ), he allowed for the full translation of the IRM source. Version 1.4 and then 1.5 were released under his guidance. By that time I had started full time at UC Davis, so my time was becoming slim catching up with all those college things. Matt graciously took up the maintainer role for IRM 1.5 and released several versions, culminating in 1.5.8-dev, the current recommended version (or 1.5.7 for the squeemish).

During all of those time, I became increasingly dissatisfied with the PHP structure that IRM was running on. Its underpinnings were still dating back from 1999, PHP 3.0 code, strange code reuse, no abstract templating, and a fixed database structure. In 2002 I attempted a partial rewrite of IRM called IRT (Information Resource Tracker) in a custom mod_perl framework (the job I had at the time was developing a custom mod_perl application for footagesearch.com – still used to this day, though its had numerous updates). It failed miserably. I can’t even find the source code.

Next up was IRM 2.0 in Python. It included some new ideas for abstract schemas and relational linking. Its asset manager actually worked. It was developed in a custom Python application server Cymbeline, which was probably a bad move as I was spending more time developing Cymbeline than IRM 2. This also fell by the wayside during those college years. But it was a great learning exercise in what worked and what didn’t.

And thats the history in a summary. If you were a contributor (I know I left a bunch out, especially when I was not active maintainer ), let me know and I’ll add you to the history

Since working on IRM, my life has changed a lot. I graduated high school (yes, IRM is that old), graduated college, and got a good “real job.” More importantly, I’ve built lots of real software in between, for personal use, research work, as well as commercial contract work. I’ve been around the block a few times, and tried on a large number of hats, all of which will reflect themselves in the design and implementation of IRM 2.

So, that gives us IRM 2, Mk III. The third incarnation of IRM2. And better, more powerful, and very different from the rest. Read on.

To begin, its still a web application. IRM is never going to be solely a desktop application, though the line between web application and desktop application is really blurring these days.

It utilizes Java. I can already hear the complaints trickling in. I’ve grown quite fond of Java at my day job, and really recognized what strengths it brings to the enterprise web application front. The Sun Java VM is now also Open Source, with the rest of the classpath libraries becoming Open Source in the near future. There are multiple Java vendors, and Java does really run everywhere these days. Java is also very quick, and the framework provided with it are top-notch. A considerably worse choice would be to move to .NET and ASP.NET.

It utilizes Enterprise JavaBeans 3.0 . These are not the EJBs of 2.0 and 1.0 days, which were, to put it mildly, horrible. EJB3.0 is actually a joy to work with, and brings with it an amazing amount of functionality, such as transactional support, security, and scalability.

It utilizes the Java Persistence API 1.0, along with Hibernate when the JPA doesn’t go far enough. Both of these technologies are wonderful to use, and greatly speed development (and makes maintenance much easier).

For the presentation layer, Apache Tapestry won out over other packages such as Struts2, JSF, and Spring. While I prefer to use Java Community standards whenever possible, I compromised for Tapestry as it is a wonderful (though not well known) package.

All of this can successfully run in the JBoss Application Server, making deployment a snap on any platform. It can also be run in any mostly-EJB3.0 compliant container, though drop-in portability will not be a focus of the first releases. By utilizing Java Persistence, we will automatically support any JDBC compliant database server, from good old MySQL, Oracle, PostgreSQL, Microsoft SQL, DB2, to the built in HSQLDB.

I am working on a master roadmap which lays out our development cycles and sets goals. Look here in the next few days for updates on how its going to look.

In short, IRM 2 will feature:

• The asset manager interface from PythonIRM2. User definable schemas for any device, item, location, or thought you can think of. Fully linkable schemas let you associate users, software, locations, peripherals, and potted plants together into an AssetMesh^tm. Since everything is associated in real life, why not let your asset manager let you express that?
• Trouble ticket system to manage user problems, your own problems, and let you keep your head on straight. Includes both user ticketing, technician ticketing, and phone support ticketing.
• User self help system, bringing back the FAQ, user trouble reporting, and other features which made IRM into such a successful integrated application for the IT helpdesk.
• And more!

To bright futures ahead, but remembering the beginnings…