Work Progressing on Army's Future Combat Systems

El_Oscuro brings us a Washington Post update on the progress of Future Combat Systems, the U.S. Army's Linux-based operating environment that has been under development for several years. The project, which currently surpasses 63 million lines of code, has received criticism for having a scope greater than that which the Army can manage. Since the program's inception, integration of commercial applications has increased the amount of code, but has also saved the developers time and money. "Boeing and the Army said they chose not to use Microsoft's proprietary software because they didn't want to be beholden to the company. Instead, they chose to develop a Linux-based operating system based on publicly available code. Boeing's Schoen said that it is designing software so that if soldiers lose their connection, the software will automatically "heal itself," retrieving the information within seconds without rebooting."

Comments

Re:And Appropriately

[Anonymous Coward]

You guys would be appalled at the bad software that's in this thing. From a bizarrely dysfunctional display system to a completely unstable and ever changing target OS. Yes, it runs linux but Boeing has decided that linux isn't good enough and is rolling out their own operating environment that we're all forced to use.

Blecch. Blecch. Blechh.

Oh, and the whole thing hinges on futuristic radios that don't work.

Yeah, I think I've been working FCS for too long. Sigh...

Re:And Appropriately

[cshotton]

Be VERY afraid. FCS/SoSCOE (System of Systems Common Operating Environment) is your worst, worst nightmare. It all squats upon an antiquated CORBA infrastructure and is the most bloated, incredibly poorly engineered PoS that has ever been birthed by an aerospace contractor. And I should know. As chief architect for the Common Operating System component of DARPA's J-UCAS program, we fought Boeing long and hard over their insistence that this architecture form the basis of the J-UCAS software infrastructure. While the idea stems from the long-running quest within the DoD to develop a true cross-service network-centric software architecture, it was built by people who completely ignored the last 15 years of lessons learned about large scale distributed systems from the Internet. It has multiple single points of failure baked into the architecture, requires outrageous amounts of RAM and CPU power to run (making it incredibly unsuitable for embedded systems use), and is licensed in such a way as to make it virtually impossible to obtain and modify without Boeing's involvement.

Furthermore, Boeing has expressed in public on several occasions that they intend for SoSCOE to make them the "Microsoft" of military systems. They are purposefully engineering a system designed to cement their position as a sole provider of OS components for network centric platforms. Nice bastardization of the open source components they are using to say the least.

Having tried repeatedly to get 2 SoSCOE nodes to communicate, we subsequently replicated 100% of the functionality that J-UCAS required using less than 150,000 lines of code and $2M of budget. Makes you wonder how long we need to support the programmer welfare for Boeing's "software engineers" and their 60 million line monstrosity if it can all be done with 400 times less code than that?