Fog Creek Software
Discussion Board




bloated old school development vs. fast new school

http://www.usatoday.com/money/covers/2002-02-07-tech-contractors.htm

But the first prototype by Raytheon was a 40-pound monstrosity, according to GAO reports, Army officials and defense industry experts. During testing, soldiers who rolled on the ground got stuck on their backs like tortoises. The helmet was so heavy, troops who were crawling couldn't lift their heads to fire rifles. A thick helmet cable got snagged in bushes so often that soldiers ripped it out.

The early Land Warrior software rarely worked, and batteries for computers and radios lasted far less than the desired 12 hours. The system failed water tests, leaking badly. During jump exercises, the bulky computer packs wouldn't fit under soldiers' parachutes.

"It was a classic example of guys sitting around a table, wishing they had this and that," says retired Army Lt. Col. Tim Eads, an analyst at the Center for Strategic & International Studies. "You ended up with a 50-pound piece of metal that soldiers hated dragging around."



===============



The Silicon Valley engineers slapped together a crude model in three months. They went to retailers Best Buy and Fry's Electronics and bought several cheap, off-the-shelf products, including Microsoft Windows CE software and a wireless card to allow Land Warrior computers to send data.

The most critical technical step: They wrote the software in common programming language used by most software engineers, rather than using old government programming language, as Raytheon had



SSSSS
Thursday, February 07, 2002

One word: "Hunger"

Raytheon is big. 100,000 employess big, I believe. They are entrenched in the Military development system. The smaller companies are just that, smaller, hungier, and trying to break into a market dominated by monster companies. And hunger can be a great motivator.

Other thoughts:
"ignored rigid Army specifications" -- it sounds like they ignored MILSPEC requirements. Those are the specifications that are meant to ensure a device works at the North Pole, on the equator, and everywhere in between. They can be extremely demanding. Was it good sense to ignore outdated, overly restrictive specs? Or did they compromise reliability in inclement weather?

The original design started around 1996, and it was probably based on pre-1995 technology. The newer designs began around 1999, and used current technology. Newer technology guarantees equal or better performance, but still lighter, more robust, and cheaper.

"The firm felt that Raytheon had followed Army specs for the project too closely. The old prototype had to be trashed and a new computer and radio system built." -- They ignored what the Army wanted and gave them what they (thought they) needed. This is risky. If the Army actually wanted what they said they wanted, then you're up a creek.

"Pacific Consultants said it could finish its prototype in six months for $2 million — more quickly and cheaply than the other bidders" -- But will they? Can they actually achieve their low bid? Or will they be yet another missed-schedule, over-budget military contractor?


I'm not trying to defend Raytheon or dismiss the smaller contractors. But I think it's important to note that this is not simply a matter of "big and slow" versus "small and fast" development methods. Raytheon played by the rules, met the military's requirements, using military-grade hardware. The smaller contractors ignored the official requirements, didn't use field-tested hardware-specs, and used untested hardware for military use. This is a paradigm shift. It could be a Good Thing, or the Army might end up with unreliable hardware that doesn't do what it wants.

Here's hoping for the former.

David Fischer
Thursday, February 07, 2002

> Raytheon played by the rules, met the military's requirements, using military-grade hardware

Yes, they couldn't care less if the product was useful.  Just a  bunch of drones biling their hours, and "fullfilling our contrctual obligations"  Useless.

SSSSS
Friday, February 08, 2002

Well, the Army brings it upon themselves for preserving oligarchies.  I don't think it's a triumph of the New School, but the fact that the military has had a need for secrecy that from time-to-time manifests in these stupidities.

Robert Milton
Friday, February 08, 2002

Interesting read.

Famous last words:
<< We fulfilled our contractual obligations and designed what the government requested, says Raytheon's Martin >>

It's not about fulfilling the contract, it's not about designing 100% according to the spec. It's all about happy customers.

Although I'm not saying to completely ignore the first two.

Jan Derk
Friday, February 08, 2002

"Yes, they couldn't care less if the product was useful. Just a bunch of drones biling their hours, and "fullfilling our contrctual obligations" Useless. "

Having spent the better part of a decade working on assorted military projects for different governments, I can tell you that meeting contractual obligations is exceedingly important.  A huge amount of time is also spent in trying to make products as useful and marketable as possible but you run into the problem that you are delivering a contract, and not a product.

One of the major problems with military contracts is that the people who assign the contract, review the design, and approve the end results are not the people who are on the ground.  What the people on the ground want and what is written into the contract rarely match. 

"It was a classic example of guys sitting around a table, wishing they had this and that," says retired Army Lt. Col. Tim Eads, an analyst at the Center for Strategic & International Studies. "You ended up with a 50-pound piece of metal that soldiers hated dragging around."

In this case, I would be willing to bet that the people sitting at the table wishing were all in military procurement and not so much a Raytheon body.

This doesn't mean that Raytheon didn't screw up the contract.  If the product had leakage and power issues, they probably did something wrong. 

At the same time, it is much more difficult to get something out the door while adhering to MIL-SPC.  The thought of using commercial solutions is alluring but prone to serious issues.  Most of the places that this equipment will be deployed is fairly inhospitable and electronics hostile.  Try running your PC without a fan at 70C for a while and see how it fares.

!
Friday, February 08, 2002

>. Most of the places that this equipment will be deployed is fairly inhospitable and electronics hostile. Try running your PC without a fan at 70C for a while and see how it fares.


Well, that should be addressed in the ANALYSIS phase.  ie:  "Should be build this damn thing in the first place ?"  Is it feasable? etc"

SSSSS
Saturday, February 09, 2002

Anyone remember the Simpson's episode where they built a car that responded to what Homer wanted ?
It feels like that episode.

Francois-Michel Larocque
Tuesday, February 12, 2002

*  Recent Topics

*  Fog Creek Home