Five years ago I started a company with a few other folks who were trying to “save” a school administration system that was written in RPG that we did not own. In this particular case, the company that owned the software and the source code was really just using the customer base as a revenue source while they pushed the customers into other school administration solutions that were not System i centric. You may have heard of PowerSchool, once owned by Apple, which is now owned by this same company. How a company that initially acquired the premier IBM Midrange (S34/S36/S38/AS/400) school administration system (once sold through IBM!) could end up selling a solution that runs on an Apple platform is a long, ugly tale that involves hundreds of folks and millions of dollars. But, the bottom line here is that IBM failed to capitalize on and captivate the mind share of system administrators that really needed a stable, reliable and easy to administer system. That failure was demonstrated in the reality that the company that acquired this software back in the 80’s missed the fact that the true asset was the system the software ran on, as well as the software itself. The hardware was the key to the solution, not just an interchangeable component of it. As the software solution was bought and sold a few more times, the beauty of the platform was lost to the need to look more “modern”.
But, that was also a bit before systems started to become commoditized. Commoditization has hurt the System i platform prospects because the system can run so many application that people miss out on the fact that running applications is only part of the solution. Running applications in a stable, reliable and easy to manage environment is the key. It is the holy grail. The System i has that. It is a slam dunk reality that no other platform, IMHO, can match.
So, now the issue for us System i die-hards is, the RPG applications of yesteryear, like the school administration solution I originally mentioned, are essentially unchanged in their 1980’s look and feel. That is precisely because of the stability of the System i. I have customers who I trained on the System 36 version of the software in 1987 that are still running the follow on version of the software on the System i today. In some cases they are using data that was keyed in on a System 36 and migrated to the System i with minimal disruption. They love this platform and, in smaller school districts, they have managed to keep their IT staff to one or two people when it comes to running the school administration solution. In one case, I consulted at a district that had NO IT staff for the System i school administration system, just a janitor who swapped the tape out of the drive while he cleaned at night. Compare this to the instructional solution arena where there might be many, many IT people needed to keep the network, application, database, web and email servers running. From a management standpoint, it is a nightmare.
So back to the “missing link”. Here is the reality of the marketplace we were facing when we started up the company:
- Loyal Midrange customers. Some had been using versions of the software for 20 years.
- Green screen stability was losing to GUI “glitz and glamor”. PowerSchool looks great! Our solution looked like “DOS”.
- The “AS/400”, although stable, was seen as a dinosaur.
Our decision, for better or for worse, was to lead with the System i. We didn’t have source code so we had to rewrite from scratch. We used the database on the System i and released the code as Open Source. We were going to focus on a web based solution rather than a plump or fat client. We initially chose Java as the programming language because we could write code that would run (theoretically) on any platform. That would alleviate the need to fight the hardware battle during the sale of a system (although we would lead with the System i in all cases). We could continue to support customers running the RPG version of the code while we delivered a web version. They could use both simultaneously since both solutions used the same database.
Five years later, how did we do?