No Data, No Model, No Technical Support, No Problem?

Note: I am preparing a letter for George Haley

I did my best to produce the desired results on schedule within budget for Thomas Lee Hull.

My best efforts were not good enough for Thomas Lee Hull.

I cannot do better than my best.

At the risk of sounding like one of those Ungrateful Whiny Cry-Baby Excuse-Making-Why-I-Can't-Do-the-Work Goof-Off Northrop Grumman Employees that management seem to see in their employees, I will attempt offer some possible reasons why I had problems doing the NXX work using the THUNDER campaign model for Thomas Lee Hull in 2008.

In 2007, I was informed that I was the Northrop Grumman "Expert" on the THUNDER Campaign model, despite having only rated myself a "3" or "Competent" on Xref for the THUNDER model, and that I was chosen to work the NXX analysis by Thomas Lee Hull.

I told Thomas Lee Hull that I required three things:
1) a current set of data sets for the the scenario in question;
2) the current street version of the THUNDER model; and
3) technical support for all of the inevitable questions and problems that come up when working with these large models

The problem is that Northrop Grumman's membership in the User Group was terminated by Air Forces Studies and Analysis years previous for some obscure political reason, our User Group fee being refunded on a pro-rated basis. This meant that NO THUNDER data, model distribution, and technical support would be available to our study efforts. Only government agencies that were User Group members, and commercial users with an active contract with a sponsoring government agency that was a member of the User Group had access to the data, model distributions, and technical support.

However we did have:
1) previously acquired THUNDER data for a long-range heavy bomber study;
2) a repaired version of a modification of the THUNDER that was provided to us for an unmanned air vehicle study by a government agency;
3) other THUNDER users at Northrop Grumman.

The NXX study was to look at future U.S.Navy requirements for a particular scenario. Our data, being several years old and created for a long-range heavy bomber study, was seriously lacking in several areas.

One OBVIOUS example of the data deficiency was that the carrier aircraft were not being permitted to fly by the THUNDER mjodel. Being a long-range heavy bomber oriented data set, there were NO air-to-air missiles required by the carrier aircraft that were not used in the long range bomber study. Modifying and adding the three air-to-air missiles to the relevant files was easy.

The other obvious example was that the thousands of targets shown in current intelligence documents were not in our data set.

What was more difficult was determining what other non-obvious deficiencies there were in the data.

Thomas Lee Hull had major heartburn when he told me to reassure him that we had "good data" as he would be briefing this work in front of admirals and did not want to look like a fool. I did not want to lie to his face by telling him that I thought the data was complete and good, BUT told him that I would not say that the data was "good".

Despite reassurances from Thomas Lee Hull that the customer was going to someday provide THUNDER data, Northrop Grumman never received any data.

We never saw the current street version of the THUNDER model.

GOOD NEWS: technical support was finally bought and paid for by Northrop Grumman.

BAD NEWS: Garth Morgan of S3I did not arrive until the second to the last week of 2008 in late December.

We briefed Garth Morgan, S3I's Vice-President and Chief Technology Officer, on the THUNDER work we had done in 2008. Garth Morgan and Wayne Zandbergen had written the code for the THUNDER model back in the 1980's.

Garth Morgan made several observations, including the following two that I consider relevant to my difficulties in producing the desired results on schedule within the budget for Thomas Lee Hull.

1) the long range heavy bomber study data we were trying to use "TOTALLY SUCKS". If Northrop Grumman were to spend large amounts of money, time, and labor, the data could be made such that it "ONLY SUCKS". At least we would then have some understanding of the data and why results "SUCKS".
2) one of the major problems that I could not solve despite much effort by myself and Christopher Vincent who was enrolled to assist me for a few weeks was figuring out why the carrier aircraft carrying a specific weapon against ground targets would "Return Before Target". The THUNDER campaign would report that these carrier aircraft would return to base before attacking its targets.

By the way, Christopher Vincent previously helped write and used regularly the ATCOM model – Advanced Tactical Combat Model. It’s the standard the Army uses for aviation requirements & training and is recognized as an Air Force standard tool as well. Google will give you more info on it. Having experience with a large model using multiple data sets, he agreed with me that I had a very difficult task.

We tried testing all of the possible reasons that would make the carrier aircraft abort its mission with no success.

Garth Morgan laughed and explained that he and Wayne Zandbergen originally wrote the THUNDER campaign model for the Air Force. They programmed THUNDER to make check that the aircraft had appropriate weapons before attacking targets. HOWEVER, the Navy weapon that we specified for the carrier aircraft to deliver against its targets is NOT considered a suitable air-to-ground target by the Air Force.

This weapons test was hard-wired into the code somewhere and not in the data; this was the reason that the carrier aircraft with that particular weapon would "Return Before Target".

IMAGINE my surprise and disappointment on learning this.


Disclaimer: SORRY - This is only what I know from my point of view. No lies, just what I see as the truth.