"You are, and you've been caught out multiple times.  You forget that some of us actually do have an involvement in military acquisition and platform assessment

http://www.strategypage.com/CuteSoft_Client/CuteEditor/Images/emteeth.gif" border="0" align="absmiddle" alt="" /> Doesn't sound irrational to me. When I was in college, the first programming course I was taught was assembly language and it taught you to be a careful and meticulous coder.  Considerations for things like code path length seem to be non existent nowadays.  I was always irked by programmers who proclaimed that a segment of code was short without consideration that if you performed it millions of times it added up. 

 

I can't speak for GF but my discussion of assembly language is partly tongue in cheek in case you missed the smiley (i.e. lighten up).  The debate of compiled versus "interpreted" code has been around for years before Java came on the scene as has the notion of simply throwing more hardware to rectify performance problems so I'm not going to rehash it here. 

 

The stability and performance problems I experienced with various Java environments had to do with the JVMs themselves and not something that could be rectified with hardware.  Badly written code can consume all the CPU cycles you throw at it. 

 

But I'm sure the capable Koreans will overcome any obstacles they come across... http://www.strategypage.com/CuteSoft_Client/CuteEditor/Images/emwink.gif" align="absmiddle" border="0" alt="" />

 

@ gf0012-aust



> You are, and you've been caught out multiple times.



Such as?



> If you want to make programmers run tight efficient code then make them do it in Assembly.



What's the point of assembly in 2009? Time is money. Recall that Koreans draw from their massive commercial electronics development to build their weapons code, where time and fast turnaround around the clock is critical for survival and delays are unacceptable. It is an entirely different mindset behind these Korean weapons engineering program.



@ sinoflex



> are used on aircraft avionics but having had experience with Java runtime environments in multi vendor environments, they can be buggy and slow.



This is why the hardware running this Korean avonics system is unusually "high-spec" and "very modern" unlike traditional embedded systems of US avionics to account for the inefficiencies of Java. In return, you get a development cycle not possible with Ada and C++.



> Personally I would feel more comfortable with a compiled language like C++ .



Performance difference between C++ native binary and JIT compiled Java classes have diminished greatly overtime, and you can always throw in more modern hardware to make up the remaining difference since the system is so portable to begin with.



The benefits of rapid development and easy code maintenance more than negates the requirement for a higher power hardware.



> My bad, ADA is the standard programming language for avionics, no?



Late comers change the rule. ADA is a dying breed even in the US weapons programming language and the standard language is C++. Koreans took an advantage of advances being a late comer and is coding their avionics and weapons in Java, because computing power is really cheap but programmer time is not.

@ Rufus



> we are talking about a multi-billion dollar high-tech program, and you think Indonesia, a country that can't even keep the handful of planes it his in its inventory airworthy, is going to buy its way in with cheap commodities?



How much is a tanker full of oil worth?

That figure is from Korean DoD's presentation to decision makers. There is a reason why Korean government likes to do business with Boeing; they are very "generous" in terms of tech transfer terms unlike Lockheed Martin.

I think you got things mixed up. The one I am talking about is Korean DoD's slide to Koreans.

 

 http://img525.imageshack.us/img525/7603/download020.jpg" />

This one shows KFX bidders and what they are bidding on. This was in back in 2007 with Saab and BAe being the msot aggressive bidders. Boeing wasn't aggressive at this time.

This slide shows the tech roll over from their indigenous army helicopter program to the next project(FA-50, which would then roll over to KFX). Koreans accumulated tech by rolling over tech from one project to another and this is a strange mixture of American, European, and Russian tech. Light blue items are government developed, deep blue items are local contractor developed, red items are purchased from foreign suppliers, yellow I don't know. According to this, their avoinics is already all local by 2008, with local computer systems running navigation, automatic self-defense, target acquisition and fire control, and mission. Sensors are a mixture of locally sourced and imported. Stuffs like radar detection and laser detection sensors are local, while others like missile warning and diversion are imported.

 

There was another slide like this that I can't find, which shows what they have accumulated for KFX from various programs and it's roughly 70%(Includes avoinics, indigenous ground targeting, RAM, shape-based RCS reduction, airframe engineering, weapons suite, etc). The AESA radar problem sourcing is taken care of with the Elta EL/M-2052 licensing deal, and the sole major omission is the engine, and the experience of actually putting together a fighter plane.

http://defencepolicy.com/data/cheditor/0808/boing_leftcore.jpg" width="640" height="359" />

This was Boeing's first KFX proposal. Korean DoD rejected it promptly and Typhoon became the leading candidate from 2007~2008, but the project stalled due to issues with bidders(Price??? Workshare???)..

 

Then something changed drastically in the early 2009, and now Boeing's revised proposal of a fully-redeveloped full-stealth 5th gen fighter based on Super Hornet layout was said to be leading, and KFX was now a definte go from several reports.